cpwf project report · 2016. 8. 2. · taking any action based on the information in this...

Research Highlights CPWF Project Report

CPWF Project Report Developing a System of Temperate and Tropical Aerobic Rice in Asia (STAR) Project Number 16 B.A.M. Bouman

International Rice Research Institute (IRRI)

for submission to the

February 2008

Contents CPWF Project Report

Page | 2

1. ACKNOWLEDGEMENTS

This report presents the main results, outcomes, and impacts of CPWF project “Developing a

System of Temperate and Tropical Aerobic Rice in Asia (STAR)” (PN16). The chapter “Project

participants” lists the key partners and their institutions. However, many partners were involved in

the project and it is impossible to name them all. Besides the key research staff, we’d like to

especially thank all the students and scholars who were involved in the project and did a lot of the

experimental work, the staff of experimental stations who took care of the experiments, the many

staff of institutes and agencies who supported project activities (local government agencies, local

stations, extension agencies, irrigation system officers), and – last but not least – the many

farmers who participated in our on-farm trials, tested aerobic rice, and subjected themselves to

“hours of interrogation” during field surveys!

Our CPWF project collaborated strongly and effectively with two research consortia and one other

project on aerobic rice development in Asia:

1. The Irrigated Rice Research Consortium (IRRC), supported largely by a grant from the

Swiss Development Agency (SDC).

2. The Consortium for Unfavorable Rice Environments (CURE), supported by different donors,

including the CPWF through CPWF project 12.

3. The project on “Developing and Disseminating Water-Saving Rice Technologies in South

Asia”, supported by a grant from the Asian Development Bank.

We held joint planning meetings and workshops, and freely shared ideas, research strategies,

protocols, and results.

Program Preface:

The Challenge Program on Water and Food (CPWF) contributes to efforts of the international

community to ensure global diversions of water to agriculture are maintained at the level of the

year 2000. It is a multi-institutional research initiative that aims to increase water productivity for

agriculture—that is, to change the way water is managed and used to meet international food

security and poverty eradication goals—in order to leave more water for other users and the

environment.

The CPWF conducts action-oriented research in nine river basins in Africa, Asia and Latin America,

focusing on crop water productivity, fisheries and aquatic ecosystems, community arrangements

for sharing water, integrated river basin management, and institutions and policies for successful

implementation of developments in the water-food-environment nexus.

Project Preface:

[**Missing]

Outcomes and Impacts CPWF Project Report

Page | 3

CPWF Project Report series:

Each report in the CPWF Project Report series is reviewed internally by CPWF staff and

researchers. The reports are published and distributed both in hard copy and electronically at

www.waterandfood.org. Reports may be copied freely and cited with due acknowledgment. Before

taking any action based on the information in this publication, readers are advised to seek expert

professional, scientific and technical advice.

Citation: Bouman BAM (2008) CPWF Project Number 16: Developing a System of Temperate and Tropical Aerobic Rice in Asia (STAR). CGIAR Challenge Program on Water and Food Project Report

series; www.waterandfood.org.


Contents

1. Acknowledgements............................................................................................... 2 2. Project highlights ................................................................................................. 7 3. Executive summary .............................................................................................. 8 4. Introduction ....................................................................................................... 12 5. Objectives .......................................................................................................... 14 6. Outcomes and Impacts ....................................................................................... 15

6.1. Objective 1: Aerobic rice varieties identified........................................................ 15 6.1.1. Method.................................................................................................. 15 6.1.2. Results China ......................................................................................... 16 6.1.3. Results India .......................................................................................... 16 6.1.4. Results Laos........................................................................................... 17 6.1.5. Results Thailand...................................................................................... 18 6.1.6. Results Philippines................................................................................... 21 6.1.7. Conclusion ............................................................................................. 21

6.2. Objective 2: Water and nutrient dynamics .......................................................... 22 6.2.1. Field experiments.................................................................................... 22 6.2.2. Method.................................................................................................. 22 6.2.3. Results China water................................................................................. 27 6.2.4. Results China water x nitrogen .................................................................. 33 6.2.5. Results China NPK................................................................................... 37 6.2.6. Results India water.................................................................................. 39 6.2.7. Results Philippines water x nitrogen............................................................ 43 6.2.8. Results Philippines nitrogen x row spacing ................................................... 49 6.2.9. Results Laos NPK .................................................................................... 54 6.2.10. Conclusion ............................................................................................. 56 6.2.10.1. Simulation modeling ................................................................................ 57 6.2.11. Method.................................................................................................. 58 6.2.12. Results .................................................................................................. 58 6.2.13. Conclusion ............................................................................................. 62

6.3. Objective 3: Sustainability ............................................................................... 63 6.3.1. Gradual yield decline................................................................................ 63 6.3.2. Yield restoration...................................................................................... 67 6.3.3. Long-term aerobic rice yields .................................................................... 70 6.3.4. Causes for yield decline............................................................................ 71 6.3.5. Yield “collapse” ....................................................................................... 77 6.3.6. Conclusion ............................................................................................. 78

6.4. Objective 4: Crop establishment ....................................................................... 80 6.4.1. Method.................................................................................................. 80 6.4.2. Results China ......................................................................................... 81 6.4.3. Results India .......................................................................................... 83


Page | 5

6.4.4. Results Philippines................................................................................... 84 6.4.5. Conclusion ............................................................................................. 86

6.5. Objective 5: Target domain.............................................................................. 86 6.5.1. Water availability and soil type .................................................................. 87 6.5.2. Target domain varieties............................................................................ 90 6.5.3. Mapping yield and water requirements........................................................ 91 6.5.4. Farmers’ practices ................................................................................... 95 6.5.5. Extrapolation domains.............................................................................100 6.5.6. Conclusion ............................................................................................101

7. International public goods ................................................................................ 103 8. Outcomes (partner change) .............................................................................. 105 9. Impacts............................................................................................................ 108

9.1. Science........................................................................................................108 9.2. Capacity building...........................................................................................108 9.3. Community impacts .......................................................................................111 9.4. Global impacts ..............................................................................................112

10. Partnership achievements................................................................................. 115 11. Recommendations ............................................................................................ 116

11.1. Policy makers ...............................................................................................116 11.2. Extension.....................................................................................................116 11.3. Research......................................................................................................117

12. Key publications ............................................................................................... 119 Published Journal papers.......................................................................................119 Journal papers submitted or in preparation ..............................................................119 Other scientific publications ...................................................................................120 Student dissertations ...........................................................................................122 Popular articles (English) ......................................................................................123 Training materials................................................................................................123

Bibliography ............................................................................................................ 124 Project participants ................................................................................................. 127

Partners with funding from CPWF ...........................................................................127 Contributing partners without funding from CPWF .....................................................128

Appendices.............................................................................................................. 130 Appendix A: Abstracts of all key publications...............................................................130 Appendix B: Ten Frequently Asked Questions and answers ............................................138 What is aerobic rice? ............................................................................................138 What are aerobic rice varieties? .............................................................................138 What is the difference between aerobic rice and upland rice? ......................................138 Why aerobic rice? ................................................................................................139 Where aerobic rice? .............................................................................................140 How to manage aerobic rice? .................................................................................140 Is aerobic rice rainfed or irrigated? .........................................................................141


Page | 6

Is aerobic rice a “mature” technology? ....................................................................142 How about aerobic rice and conservation agriculture? ................................................142 How sustainable is aerobic rice? .............................................................................143

Appendix C: Detailed recommendations for further research..........................................144

Project Highlights CPWF Project Report

2. PROJECT HIGHLIGHTS

The project “Developing a System of Temperate and Tropical Aerobic Rice in Asia (STAR)

undertook strategic research to develop sustainable aerobic rice systems for water-scarce irrigated

and rainfed environments in Asia. Aerobic rice is a production system in which specially developed

rice varieties are grown in nonsaturated soils without ponded water just like wheat or maize. The

target environments are areas where water is too short to grow conventional lowland rice, either

rainfed or supplementary irrigated. In the Yellow River Basin of China, with a temperate climate,

we have demonstrated that aerobic rice yields of 6 t ha-1 are attainable with about half of the

water needed to grow lowland rice. In average rainfall years, farmers would need to give only 2-3

supplemental irrigations. The profitability is comparable with that of other food crops such as

maize and soybean, depending on (yearly fluctuating) relative commodity prices (sometime

profitability is lower, sometimes higher). Farmers like aerobic rice because it contributes to food

self-sufficiency and requires less labor than transplanted lowland rice. It also allows them to

diversify their cropping system. Moreover, aerobic rice can stand flooding and is an ideal crop for

the large areas that get annually flooded by heavy rainfall or overflowing rivers that destroy the

other crops. In the tropics, the development of aerobic rice is less advanced. In central India, in

the Indo-Gangetic Plain, we identified rice varieties that can be grown in aerobic conditions,

producing 4-4.5 t ha-1 and using 30-40% less water than lowland rice at the same yield level. In

the Philippines, although yield potentials of 6 t ha-1 have been demonstrated, attainable yield

ranged from 2.9 to 3.8 t ha-1 in the dry season, and from 3.9 to 4.5 t ha-1 in the wet season. A risk

of yield decline was demonstrated at a few sites caused by soil-borne pests (such as nematodes),

nutrient disorders, or a combination of both. In our sites in Northeast Thailand and Laos, breeding

lines were identified with yield potentials of 2 (Thailand) to 3.5 (Laos) t ha-1. Further research and

development is needed to bring tropical aerobic rice to fruition, mainly on variety improvement

(increasing yield potential and adaptation to aerobic soil) and sustainability. In conclusion, aerobic

rice holds promise for those farmers in water-short irrigated or rainfed environments where water

availability at the farm level is too low, or where water is too expensive, to grow flooded lowland

rice.

Executive Summary CPWF Project Report

3. EXECUTIVE SUMMARY

Background, objective, methods

New technologies need to be developed to assist farmers to cope with water shortages in rice

production. Aerobic rice is a production system in which specially developed, input-response rice

varieties with “aerobic adaptation” are grown in well-drained, nonpuddled, and nonsaturated soils

without ponded water. Evidence of feasibility comes from northern China, where breeders have

produced first-generation (temperate) aerobic rice varieties that use only 50% of the water used in

lowland rice. However, sustainable crop-soil-water management recommendations are lacking. A

shift from continuously flooded to aerobic conditions may have profound effects on sustainability

(e.g., soil health), which needs study to develop sustainable production systems. Moreover, there

are no aerobic rice varieties for the tropics. In this CPWF project, strategic research was

undertaken to develop sustainable aerobic rice systems for water-scarce irrigated and rainfed

environments in Asia. The objectives were to:

1. Identify and develop aerobic rice varieties with high yield potential

2. Develop insights into key processes of water and nutrient dynamics

3. Identify key sustainability issues, and propose remedial measures

4. Develop practical technologies for crop establishment

5. Characterize and identify target domains

Project partners and sites were located in the Philippines and in three CPWF benchmark basins:

Yellow River, Indo-Ganges, and Mekong. Research methodologies included pot experiments, on-

station and on-farm field experiments, simulation modeling, GIS, and farmer surveys.

Results and conclusions

Breeding of aerobic rice varieties is most developed in China, and varieties were identified with

demonstrated yield potentials of 6 t ha-1 (HD502, HD297), in relatively dry soil with soil water

tension going out of the 100 kPa measurement range. In India, tropical aerobic rice varieties with

4.5 t ha-1 yield potential were identified (Pusa Rice Hybrid 10, Proagro6111, Pusa834, Apo

(PSBRc9), and in the Philippines of 6 t ha-1 yield potential (Apo, UPLRi5, Magat). However, these

varieties had a relatively lower tolerance to dry soil conditions than the Chinese varieties in that

soil water tensions had to stay below 30-40 kPa to reach these high yields. Under rainfed

conditions in Laos and NE Thailand, breeding lines were identified with yield potentials of 2 (NE

Thailand) to 3.5 (Laos) t ha-1.

At Beijing, maximum aerobic rice yields of HD297 in controlled field experiments were 5-5.6 t ha-1

with 600-700 mm total irrigation plus rainfall water. Yields were 2.5-4.3 t ha-1 with 450-550 mm

water input, while yields dropped to 0.5 t ha-1 in very dry soil (water tension higher than 100 kPa)

around flowering, which increased spikelet sterility. Irrigation is essential at flowering time if there


Page | 9

is no rain. The average seasonal evapotranspiration (ET) requirement was 600 mm. With shallow

groundwater, capillary rise can meet most of the ET and there may be no need to irrigate. With

deep groundwater tables, the net irrigation needs are 167 mm in a typical ‘wet rainfall year” (2

times irrigation), 246 mm in a typical “average rainfall year” (3 times irrigation), and 395 mm in a

typical “dry rainfall year” (4-5 times). Simulations showed that, on typical freely-draining soils of

the YRB, aerobic rice yields with HD297 can reach 6 t ha-1, with 477 mm rainfall and 112-320 mm

irrigation water. The application of any amount of fertilizer N either reduced yield or kept yield at

the same level as without N fertilization. Farmers’ fields in northern China may have been over

fertilized for many years to the extent that they are now ‘saturated’ with N. Moreover, they may

receive large amounts of N through atmospheric deposition. Nitrogen omission in aerobic rice-

wheat cropping systems in Mencheng County caused a marked decline in yield of both aerobic rice

and winter wheat, whereas P and K omission had less effect. P and K became yield limiting for

winter wheat, because of a higher demand for nutrients by winter wheat than aerobic rice.

At Delhi, aerobic rice in field experiments yielded more than 4 t ha-1 when irrigated at 40

kPa soil water tension in the root zone. Compared with typical amounts of water applied to lowland

rice fields (under alternate wetting and drying), the amounts applied in the aerobic rice

experiments of 780-1324 mm (irrigation plus effective rainfall) translated into 30-40% water

savings for production levels of 4-4.5 t ha-1. The fertilizer application rates in the experiments were

150 kg N ha-1, 60 kg P ha-1, and 40 kg K ha-1.

In the Philippines, maximum experimental yields of variety Apo in the dry season ranged

from 2.9 to 3.8 t ha-1, with the exception of 1 site/year when yields went up to 5.4-6.1 t ha-1.

There was hardly any effect of irrigation water application rate because at most sites, shallow

perched water tables developed during the experiments that reached up and into the root zone,

keeping soil water tensions within 0-30 kPa. Therefore, high yields were realized with as little as

274-590 mm total water input. Yield responded positively to fertilizer N applications, with an

application of 120 kg ha-1 sufficient to reach maximum yields. In the wet season, soil water

tensions in the root zone were mostly between 0-15 kPa, because of heavy rainfall and shallow

groundwater tables. Maximum yields were 3.9-4.5 t ha-1. Fertilizer N addition increased yields, and

the amount needed for high yields varied from 60 to 120 kg N ha-1 across sites. Crops lodged at

some sites with 120 and 150 kg N ha-1 because of heavy winds and strong rains. The risk of

lodging needs to be included in formulating fertilizer N recommendations.

The few experimental results in Laos (1 year, 2 sites) show a large variability in nutrient

response among sites.

In the long-term aerobic rice experiment at IRRI, yield of Apo under continuous aerobic conditions

declined (relative to flooded conditions) during the first 10 seasons, but increased again afterwards

(in the dry season). The yield in the same year in a “new” aerobic field (after continuous flooded

cropping) was 2.5 t ha-1 higher than in a seventh-season continuously cropped aerobic field. Levels

of the Meloidogyne graminicola nematode were high, but the typical patchiness related to the

occurrence of nematodes was not observed. The yield decline could be reversed by crop rotations

(two seasons), fallowing (two seasons), and flooding (three seasons). Yield decline could not be


Page | 10

reversed by the application of micro-nutrients, P, or K. However, crop growth was consistently

improved by the application of N fertilizer in the form of urea or Ammonium sulfate, with the latter

being much more effective than the first. Beside the gradual yield decline under continuous

cropping of aerobic rice, we encountered two cases of “immediate yield collapse” in fields cropped

to aerobic rice for the very first time in the Philippines. Despite large amounts of irrigation water

input and large doses of fertilizer N, yields failed completely. Preliminary evidence of genotypic

variation in response to “soil sickness” was found.

Direct dry-seeded aerobic rice is not very responsive to row spacing and seed rate (within the

limits tested). Row spacings between 25 and 35 cm, and seeding rates between 60 and 135 kg ha-

1 generally gave the same yields in China and the Philippines. In India yields declined fast with

seeding rates below 40 kg ha-1. In practice, the “unresponsiveness” of aerobic rice to seed rate

and row spacing means that farmers are rather flexible in choosing their own seed rates and

spacings.

Aerobic rice holds promise for farmers in water-short irrigated or rainfed environments where

water availability at the farm level is too low, or where water is too expensive, to grow flooded

lowland rice. When the percolation losses in flooded rice are 3.5 mm d-1 or higher, aerobic systems

with flash-flood irrigation will require less water, and if the percolation losses are 0.5 mm d-1 or

lower, only aerobic systems with sprinkler irrigation require less water. In northern China, the

target areas are where water availability is 400-900 mm during the cropping season. In the central

part of the YRB, yields of 5-6 t ha-1 are attainable with 0-220 mm irrigation application (in average

rainfall years; groundwater 2 m deep or less). In these regions, farmers who currently try out

aerobic rice attain usually 3-4 t ha-1, with 150-220 mm supplementary irrigation. With these

yields, financial returns to aerobic rice can be more or less than to upland crops (maize, soybean),

depending on relative market prices (that fluctuate among years). Reasons for adoption are:

having own rice on the farm, ease of establishment and less labor needs (compared with lowland

rice), good eating quality. Negative views are: low yields, difficult to control weeds, insufficient

extension support, difficult to market. Aerobic rice is unique in its characteristics to withstand both

flooding and dry soil conditions, which make it an ideal crop for areas prone to surface flooding

where other crops would suffer or fail.

Impacts

Four international journal papers have been published, seven are submitted or in advanced stage

of preparation, three international journal papers include parts of our CPWF results, and twelve

national journal papers, book or proceeding chapters are published. More than 20 posters have

been presented, and oral presentations made at workshops, conferences, and other fora. We

organized an international Aerobic Rice workshop in Beijing, October 22-25, 2007, which attracted

about 100 participants.


Page | 11

Twenty-two graduate and undergraduate students were involved in the project, and most

of them have completed their theses. Aerobic rice was part of 32 training courses on water

management in rice production systems organized conjunctively by the Water Workgroup of the

Irrigated Rice Research Consortium and the CPWF project. In total, an estimated 1589

professionals received training on aerobic rice. Through farmer school days and field

demonstrations, an estimated 1875-3750 farmers were reached with information about aerobic

rice.

The major impact of aerobic rice at farm household level is that it offers farmers the choice

to grow rice when water is too scarce to grow lowland rice, and allows them to diversify their

cropping system which will augment resilience and increase overall sustainability. Adoption will

(among others) depend on relative market prices the different crops available to farmers are

expected to fetch in a particular year, the degree of water scarcity, availability of quality seeds and

extension support, and – last but not least – farmers’ preference to grow their own rice for self

sufficiency rather than depending on markets.

Extrapolation domain construction and scenario analysis (by CPWF-Impact Study) suggest

that aerobic rice can have large impacts in India and countries in the lower parts of the Mekong

Basin. Most of the impacts would occur from adoption in India and Thailand and would include

increased yields, production, and area grown to rice, reduced rice prices, and reduced levels of

child malnutrition. The IMPACT-WATER model founds that India will be a net importer of rice in

2050. Adoption of aerobic rice would reduce the amount of required imports, which could be very

important if climate change stresses rice-growing areas in the tropics and reduces the amount of

rice available for trade.

Key recommendations

The aerobic rice technology can be considered sufficiently mature for dissemination in the Yellow

River Basin (YRB), but not so for most of tropical Asia where more R&D is needed to create

sustainable and high-yielding aerobic rice systems. “Aerobic rice” should be recognized as a special

crop type besides “lowland rice and “upland rice” to facilitate the collection of statistics on adoption

and spread of aerobic rice, and to encourage extension agencies to address aerobic rice. In the

YRB, local and national governments should strengthen the formal and informal extension system

at all levels. In most of tropical Asia, preference should be given to setting up dedicated aerobic

rice breeding programs and strengthening the R&D capacity to develop sustainable production

systems. Further research should focus on increasing the yield potential and attainable yield at

farm level through breeding and improved management (especially nutrients), quantifying real

evapotranspiration rates and extrapolating field-level water savings to regional scale, inventorizing

soil health issues in the target domains, establishing long-term continuous cropping experiments

to address sustainability issues, understanding and solving the phenomenon of yield collapse as

observed at some sites in the Philippines, and understanding the biophysical and socio-economic

factors that lead to adoption by farmers

Introduction CPWF Project Report

4. INTRODUCTION

Rice is the staple food in Asia but also the single biggest “user” of fresh water. It is mostly grown

under submerged soil conditions and requires much water compared with other crops. The

declining availability and increasing costs of water threaten the traditional way of irrigated rice

production. Moreover, lack of rainfall is a major production constraint in rainfed areas where many

poor rice farmers live. An efficient use of water is critical to help reduce poverty and safeguard

food security in water-scarce areas in Asia.

Water requirements can be lowered by reducing water losses by seepage, percolation, and

evaporation. Promising technologies include saturated soil culture, intermittent irrigation, and the

system of rice intensification. However, these technologies still use prolonged periods of flooding,

so water losses remain high. A fundamentally different approach is to grow rice like an upland

crop, such as wheat, on non-flooded aerobic soils, thereby eliminating continuous seepage and

percolation and greatly reducing evaporation. Traditional upland rice has been bred for unfavorable

uplands to give a stable though low yield under minimal external inputs. Previous experiments of

growing high-yielding lowland rice under aerobic conditions have shown great potential to save

water, but with a severe yield penalty. A new type of rice is needed to achieve high yields under

high-input aerobic conditions.

Evidence of feasibility comes from northern China, where breeders have produced first-

generation (temperate) aerobic rice varieties with a high yield potential using less water than

lowland rice. However, initial high yields are difficult to sustain and yields may decline after 3–4

years of continuous cropping. There are no aerobic rice varieties for the tropics and crop-soil-water

management recommendations are lacking. A shift from continuously flooded to aerobic conditions

may have profound effects on sustainability (e.g., soil health). We also need to deepen our

understanding of the potential target domains along with the biophysical and socioeconomic

circumstances of the farmer beneficiaries.

In this CPWF project, strategic research was undertaken to develop sustainable aerobic

rice systems for water-scarce irrigated and rainfed environments in Asia. The project lasted from

October 2004 to March 2008. Project partners and sites were located in different countries in three

of the CPWF’s benchmark basins and in the Philippines (Table 4.1; chapter “Project participants”).

This report presents the results, outputs, outcomes, and (initial) impacts. The results and outputs

are reported by objective (see next chapter). We also specified the activities undertaken, and

results achieved, by country to easily extract the outputs per benchmark basin.


Page | 13

Table 4.1. Country partners by CPWF benchmark basin and relative weight of activities.

CPWF benchmark basin Project partner country Weight (%)

Yellow River China 45

Mekong Laos, Thailand 5

Indo-Gangetic India 20

Other Philippines 30

Objectives CPWF Project Report

5. OBJECTIVES

The goal is to ensure food security and increase sustainable livelihoods in rural and urban Asia, by

easing water scarcity as a constraint to agriculture, economic development, and nature

conservation. The objectives are to develop prototype aerobic rice production systems for water-

scarce environments, by which farmers grow rice as a dry-field crop in irrigated, or rainfed, but

non-flooded fields, achieving high yields while sharply reducing their water use and non-productive

outflows. Specifically, to

1. Identify and develop aerobic rice varieties with high yield potential (Aerobic rice varieties

identified)

2. Develop insights into key processes of water and nutrient dynamics that allow us to derive

prototype management practices (water and nutrient dynamics).

3. Identify key sustainability and environmental impact issues, and propose remedial

measures, such as crop rotations (sustainability)

4. Develop practical technologies for crop establishment (originally, “weed control” was

included but dropped in favor of more focus on establishment)

5. Characterize and identify target domains and quantify the potential amounts of water

savings and rice production (target domain)

In China, aerobic rice varieties have been developed since the mid-eighties and emphasis was put

on objectives 2-5. In the Philippines, IRRI runs an aerobic rice breeding program and, like in

China, emphasis was put on objectives 2-5. In India, variety screening for aerobic conditions was

initiated in 2001. We continued variety screening (objective 1) and moved into objectives 2 and 4

by combining variety screening with water and seed rate experiments. In Thailand and Laos, no

screening for aerobic conditions was done before, and we focused mainly on objective 1 with some

on-farm testing with different management conditions in Thailand (objective 4).


Page | 15

6. OUTCOMES AND IMPACTS

All field experiments and screening trials had 3 or 4 (mostly) replications, and statistical analyses

are provided as much as possible (in a few cases only, no statistical analyses were reported by

project partners). Throughout the report, all yields are expressed at 14% moisture content, and

refer to rough rice (nunhusked and unmilled). Total biomass is expressed at 0% moisture content

after oven-drying. In the method sections, we summarize the treatments and site descriptions on

main lines only for brevity sake; full experimental details are available on request or are reported

in journal papers listed at the end.

6.1. Objective 1: Aerobic rice varieties identified

6.1.1. Method

Actual breeding of aerobic rice varieties was not part of the project, but we “tapped’ into ongoing

breeding activities and focused on selection and evaluation of breeding lines and varieties at our

project sites in potential target domains. We participated in advanced breeding trials and multi-

location testing, and conducted our own field experiments and participatory variety selection

(Table 6.1.1).

Table 6.1.1. Sites of germplasm selection activities per country

Activity/country Philippines China India Laos Thailand

Advanced

breeding trials

Los

Banos,

Dapdap

Various

locations

in N

China

Sanasomboun,

Phonethong,

Saythany

Multi-variety

field

experiments

Munoz,

San

Ildefonso,

Dapdap,

Los Banos

Delhi Udon Thani,

Khon Kaen,

Ubon

Rachathani

Farmer-

participatory

selection

San

Ildefonso

Secunderabad

(Bulandshahar)

Nong Khai,

Mahasarakham,

Phimai

For each country or site, either released varieties or breeding lines were identified with best local

suitable were identified


Page | 16

6.1.2. Results China

CAU runs a special aerobic rice breeding program and we used promising materials for testing at

our project sites A large number of varieties are identified for northern China, depending on local

climate and cropping systems. Between 1986 and 2005, 58 aerobic rice varieties have been

released (though classified as “improved upland rice”). In our experiments, yields of up to 6 t ha-1

were obtained with HD502 and HD297 in aerobic soil with soil water potentials at 20 cm depth

going beyond 100 kPa.

6.1.3. Results India

We summarized the results of screening trials of several promising lowland and upland

genotypes/hybrids from India and IRRI under soil water conditions kept close to field capacity in

2001-2004. Rice genotypes Anjali, Annada, IR55419-04, Magat, Proagro 6111, RR 165-1160, RR

51-1, Pusa Rice Hybrid 10, Proagro6111 (hybrid), Pusa834 and IR55423-01 (Apo1) yielded more

than 4.5 t ha-1 (Figure 6.1.1).

Figure 6.1.1. Yield of highest-yielding genotypes under aerobic soil conditions kept close to field

capacity, IARI-WTC station, Delhi, 2001-2004.

In 2005-2006, we extended the variety testing to three different soil water regimes (soil close to

saturation, irrigation at 20 kPa soil water tension in the root zone, and irrigation at 40 kPa soil

water tension in the root zone) and included new genotypes. The details of these trails are

reported in the paragraph “Water and nutrient dynamics” below. We concluded that rice genotypes

0.0

1.0

2.0

3.0

4.0

5.0

6.0

Sharbati

RR 354-1

IR 77298-5-6

IR 73005-23-1-3-3

RR 348-6

CT 6510-24-1-2

IR 55419-04

IR 71604-4-1-4-7-10-2-1-3

IR 72875-94-3-3-2

ADITYA

IR 69715-72-1-3

AR 166-645

ANJALI

ANNADA

RR 51-1

RR 165-1160

MAGAT

Proagro 6111

Yie

ld (

t ha-1

)


Page | 17

Pusa Rice Hybrid 10, Proagro6111 (hybrid), Pusa834 and IR55423-01 (Apo1) were highly tolerant

to aerobic conditions, as they produced grain yields of more then 4 t ha-1 under aerobic production

system irrigated at 40kPa soil moisture tension in both the seasons (Kharif 2005 and Kharif 2006).

In 2005-2006, we also tested genotype in farmer Participatory variety selection (PVS)

trials in Secunderabad, Bulandshahar. The hydrologic conditions such as soil water tension and

groundwater table depth were not recorded. An example of results obtained in 2006 is given in

Figure 6.1.2. The yields of the tested rice genotypes under aerobic rice conditions were similar to

yields obtained by farmers in the region using transplanted rice under conditions of alternate

wetting and drying. Among the 15 varieties evaluated in the PVS trial, farmers selected Proagro

6111, Pusa Rice Hybrid 10, Apo (IR55423-01) and Pusa Sugandh 3 based on their yield under

aerobic conditions, grain fineness, and higher market price.

Figure 6.1.2. Yield of various genotypes grown in farmers’ fields (PVS trials) under aerobic

conditions, Secunderabad, Bulandshahar, 2006.

6.1.4. Results Laos

Screening of varieties suitable for aerobic conditions started in Laos with the STAR project in 2005.

Screening trials were carried out in two environmental conditions: one site in the upland rainfed

environment and two other sites on upper toposequence positions in rainfed lowland areas with

uncertain water supply (both sites rainfed). A number of 16 to 24 genotypes were selected from a

drought prone screening program, eight of them were traditional varieties selected from upland

condition, whereas the others were promising new lines including two improved varieties TDK3 and

TDK7 and two traditional varieties Hom1 and Kam11. Four lines, IR55423-0166, B6144F-MR-6-0-

0, TDK10021-B-24-19-1-B, andTDK10047-2B-6-1-1, recorded yields in excess of 3 t ha-1 under,

0.0

1.0

2.0

3.0

4.0

5.0

6.0

Pus

a R

H10

Inda

m 1

00 0

01

Proa

gro

6111

Sarb

hati

IR55

419-

04

IR74

371-

54-1

-1

IR74

371-

46-1

-1

IR55

423-

01

Nav

een

Pusa

Sug

andh

3

Pusa

Sug

andh

4

Pusa

Sug

andh

5

Van

dana

Ann

ada

Pusa

834

Pad

dy Y

ield

(t h

a-1)


Page | 18

while B6144F-MR-6-0-0 had a constant and significantly superior yield performance of 3.6 t ha-1 at

all three locations. Figure 6.1.3 gives the yield measured in 2006. No soil water tensions were

measured in the screening trials.

Figure 6.1.3. Yields of different genotypes tested at three locations in Laos, 2006 wet season.

6.1.5. Results Thailand

Screening of varieties suitable for aerobic conditions started in North-East Thailand with the STAR

project in 2005. Promising local improved upland varieties and aerobic rice varieties from IRRI

were tested at a number of target locations in NE Thailand on-farm and on-station throughout

2005-2007. Most materials were tested under rainfed conditions and a large variability in yields

was obtained across sites and years. No soil water tensions were measured in the screening trials.

Like in Laos where screening for aerobic rice varieties started only recently, yields were much

lower than in China, Philippines, or India, Tables 6.1.1-6.1.3. Under flooded conditions, yields

could go up to 5 t ha-1, but they dropped to 2 t ha-1 and below under most rainfed conditions (even

with 2074 mm at Nong Khai in 2005, Table 6.1.1).

Average grain yield of aerobic rice for rainfed low land 2006 in Laos

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

1. IR

5542

3-01

66

2. B

6144

F-M

R-6-0

-0

3. T

DK10021

-B-2

4-19

-1-B

4. T

DK10047

-2B-6

-1-1

5. IR

7437

1-70

-1-1

6. IR

1501

2-19

-1-1

7. IR

6890

8-B-7

8-2-

2-B-1

8. IR

7405

3-14

4-2-

3

9. IR

7300

5-23

-1-3

-3

10. C

hao den

g

11. M

addev

e

12. M

akfa

i

13. I

R7152

5-19

-1-1

14. T

DK1003

7-B-2

1-1-

2-1-

1

15. H

om1

16. k

am1

(Check

)

Yie

ld k

g/h

a

Yield kg/ha

Sanasomboun

Yield kg/ha

Phonethong

Yield kg/ha

Saythany


Page | 19

Table 6.1.1. Grain yield of various cultivars under rainfed conditions in farmers’ fields at Nong

Khai, Mahasarakham, and Phimai, Thailand, 2005. Total seasonal rainfall was 2074 mm at Nong

Khai, 711 mm at Mahasarakham, and 479 mm at Phimai.

Grain yield (t ha-1)

Cultivar Nong

Khai

Mahasarakham Phimai

SKN 1.1 2.0 0.5

SPT1 0.8 1.0 0.2

RD10 1.0 1.3 0.1

SMJ 0.6 1.9 1.1

UBN92110 - 1.1 -

UBN91051 - - 0.3

Rai Nonsung - 1.7 -

RD15 - - 0.3

HY71 1.4 - -

Irrad SMJ 1.2 - -

Local var - 1.8 0.0

Mean 1.0 1.5 0.4

SED 0.2 0.1 0.1


Page | 20

Table 6.1.2. Grain yield and days to flowering (DTF) among rice lines, at Udonthani Rice Research

Center (UDN) and Khon Kaen Rice Research Center, Thailand, 2005. Seasonal rainfall was 627 mm

at Khon Kaen.

Grain yield (t ha-1)

UDN (flooded) KKN (Rainfed)

Cultivar Grain yield

(t/ha)

DTF

(days)

Grain yield

(t/ha)

DTF

(days)

SKN 3.80 87 0.83 102

SPT1 5.10 96 0.30 121

RD10 5.10 96 0.34 120

SMJ 2.40 77 0.79 103

Standard error 0.20 0.3 0.06 0.7


Page | 21

Table 6.1.3. Yield and days to flowering of different genotype groups, evaluated under three

growing conditions (flooded, rainfed, and flooded with drought at flowering) at Ubon Ratchathani

Rice Research Center, Thailand, 2006.

Grain yield (t ha-1) Days to flowering (days)

Genotype group Flooded Rainfed Drought Flooded Rainfed Drought

Rainfed (IRRI) 2.3 2.2 1.3 90 87 95

Irrigated (IRRI) 2.3 2.2 1.3 91 83 96

Aerobic (IRRI) 2.1 2.3 1.7 82 75 86

Upland (IRRI) 1 1.9 1.7 0.9 88 86 97

Upland (IRRI) 2 2.1 2.1 1.3 89 80 94

Aerobic Med 2.1 2.1 1.5 86 79 90

RF (Thai) 1.7 2.0 1.2 84 86 89

Upland (Thai) 1.5 1.7 1.7 84 91 88

Regional check 1.7 2.0 1.2 84 91 88

Mean 2.0 2.0 1.3 86 84 91

Max 2.3 2.3 1.7 91 91 97

Min 1.5 1.7 0.9 82 75 86

lsd0.05 0.2 0.2 0.1 1 1 1

CV 24.40 21.54 22.05 3.90 3.35 2.29

6.1.6. Results Philippines

IRRI runs a special aerobic rice breeding program and we used promising materials for testing at

our project sites. Under aerobic soil condition with irrigation applied at thresholds of 20-40 kPa soil

water tension at 20 cm depth, the following varieties have yield potentials of around 5-6 t ha-1:

“Apo” (PSBRc9), UPLRI5, PsBRc80. Sporadically, yields of more than 6 t ha-1 have been obtained

with soil water tensions below 20 kPa. The hybrid variety Magat is liked by farmers but found too

expensive (seed costs).

6.1.7. Conclusion

Breeding of aerobic rice varieties is most developed in China where aerobic rice varieties have

been released since 1985. Although yield potentials of 7-8 t ha-1 are mentioned in various Chinese

sources (Wang Huaqi, 2002), we obtained maximum yields of 6 t ha-1 in the STAR project under

well defined experimental conditions. The tested varieties (HD502, HD297) tolerated relatively dry

soil conditions with soil water tension going out of the 100 kPa measurement range. The Chinese


Page | 22

varieties being released for the N China Plain are not suitable for tropical conditions such as found

in the Philippines, Laos, and Thailand, but may be suitable for northern India. In the Philippines,

tropical aerobic rice varieties with the same yield potential (around 6 t ha-1) as in N China have

been identified, but with much lower tolerance to dry soil conditions (i.e., water tensions had to

stay below 40 kPa to reach the high yields). Further breeding efforts need to focus on increasing

the tolerance to dry soil conditions while maintaining yield potential. Under similar soil water

conditions as in the Philippines, screening of varieties in India has resulted in the identification of

germplasm (not all stable varieties yet) with a yield potential of 4-5 t ha-1. Under rainfed

conditions in Laos and NE Thailand, first screening of germplasm (without a dedicated aerobic rice

breeding program to choose from) has identified material with yield potentials of 2 (NE Thailand)

to 3.5 (Laos) t ha-1 yields under rainfed or irrigated aerobic conditions.

6.2. Objective 2: Water and nutrient dynamics

Water and nutrient dynamics were studied in on-station and on-farm field experiments and with

simulation modeling. Here, we report on these activities separately. Following Tuong et al. (2005),

we computed crop water productivity with respect to combined irrigation water input and rainfall

(WPIR):

Grain yield WPIR =

Irrigation + rainfall (kg grain m-3 water)

6.2.1. Field experiments

6.2.2. Method

We analyzed and synthesized field experiments we performed in the years before the STAR project

actually started, and performed new experiments during the STAR project itself. Table 6.2.1 gives

an overview of the location of the experiments.


Page | 23

Table 6.2.1. Sites of water and nutrient field experiments per country

Activity Philippines China India Laos Thai

Water, N

Water x N

San Ildefonso,

Munoz

Beijing:

Changping,

Shanzhuang

WTC-IARI

station, Delhi

N x row

spacing

San Ildefonso,

Dapdap, Munoz

N,P,K Shuanghu

(Mencheng)

Houay Kot,

Somsanouk

In China, water and nitrogen (N) experiments done near Beijing were synthesized for the period

2001-2006. In all these experiments, the aerobic rice variety HD297 was direct dry seeded, with

3-5 irrigation water treatments aiming at different amounts applied and distribution over the

cropping season. In Changping, we had irrigation water treatments in 2001-2002, and irrigation by

various N treatments in 2003-2004 (N details in Table 6.2.2). In Shangzhuang, we had 3 irrigation

by 3 N treatments in 2005-2006. The N treatments were 0 (N0), 75 (N1) and 150 kg fertilizer-N

ha-1 (N3). The soil at Changping was sandy (80% sand, 13% silt, 7% clay) down to 2 m, with a

groundwater table deeper than 20 m. The soil at Shangzhuang was more loamy (60% sand, 30%

silt, 10% clay) down to 1 m, with groundwater depth fluctuating between 0 and 1.2 m. At all sites

in China, irrigation water was applied by flash-flooding. At Shangzhuang, we introduced a third

treatment factor in 2006: seeding rate. Whereas in 2005, the seeding rate was 135 kg ha-1, in

2006 we used 135 kg ha-1 in the treatment combinations of W1 and W3 by N0 and N2, whereas we

used a reduced seeding rate of 65 kg ha-1 in all water x N treatment combinations. Full

experimental details of the Beijing experiments are given by Yang Xiaoguang et al. (2005),

Bouman et al. (2006), Xue et al. (submitted, 2007), and Limeng Zhang et al (in prep 2008).

We computed evapotranspiration rates from the soil water balance in the wettest soil

water treatments, using measured data for inputs (irrigation, rainfall) and changes in soil water

content from installed neutron probes, for our field experiments in Changping (we did not do this

for Shangzhuang because of difficulties of estimating capillary rise from the shallow groundwater

contribution to the soil water balance). We then computed net irrigation water requirements as

seasonal evapotranspiration minus rainfall, for three types of rainfall years (wet, medium, dry)

derived from 50 years of weather data at Changping.


Page | 24

Table 6.2.2. Fertilizer N rates in field experiments at Changping, 2002-2004.

Year Fertilizer N application (kg ha-1)

2002 none

2003; exp 1 0

113

150

2003; exp 2 225/3splits

225/5splits

2004 0

75

125

175

225

In 2005-2006, nutrient experiments on nitrogen (N), phosphorus (P), and potassium (K) were

undertaken in rice-wheat and wheat-rice rotations in a farmer’s field in Shuanghu village (33°17′

N, 116°33′ E, 25 m asl), Mengcheng County, Anhui Province. The soil is classified as Fluvisol

derived from river sediments, with 10% sand, 58% silt, and 32% clay. Groundwater depth was

unknown at the start of the experiments. The cropping sequences were aerobic rice - winter wheat

(AR-WW) from June 2005 to June 2006 and winter wheat - aerobic rice (WW-AR) from October

2005 to October 2006. Thus, the two aerobic crops were grown in succeeding years (2005 and

2006), while the two winter wheat crops were present during the same winter season

(2005/2006). Fertilizer treatments were an optimal dose of N, P, and K (NPK), of P and K (PK, N

omission), of N and K (NK, P omission), of N and P (NP, K omission), and of a control (CK) (Table

6.2.3). Aerobic rice, cv. Han Dao 502, was drilled at a rate of 45 kg ha-1 in 2005. Because of the

low emergence rate in 2005 we increased the seed rate to 75 kg ha-1 in 2006. A surface irrigation

of 60 mm was applied by flash flooding on June 16, June 22, and August 3 in 2005, and on June

11, June 17, August 8, and August 22 in 2006. Winter wheat (cv. Yumai 18) was sown on 23

October 2005 at a rate of 187.5 kg ha-1. Full experimental details are given by Xiao Qindai et al.

(in prep 2008).


Page | 25

Table 6.2.3. Fertilizer rates (kg ha-1) applied to aerobic rice and winter wheat at Shuanghu, 2005

and 2006.

Treatments Aerobic rice Winter wheat

N P2O5 K2O N P2O5 K2O

NPK 120 50 75 180 50 90

PK 0 50 75 0 50 90

NK 120 0 75 180 0 90

NP 120 50 0 180 50 0

Control 0 0 0 0 0 0

In India, field experiments focused on variety evaluation under different water regimes at the

station of the IARI-Water Technology Centre at Delhi. In total, 12 varieties were screened in 2005,

and 24 in 2006 and 2007. The three irrigation regimes were: irrigation water applied (through

flash flooding) when soil water tension at 20 cm depth reached “0 kPa” which means keeping the

soil close to saturation (I0), when soil water tension reached 20 kPa (I20), and when soil water

tension reached 40 kPa (I40). The soil in the top 30 cm is a clay loam with 36% sand, 35% silt,

and 29% clay. Groundwater is deeper than 2 m. In all experiments, the aerobic rice varieties were

dry seeded at the rate of 80 kg ha-1.

In the Philippines, we had water x N field experiments in the dry seasons of 2005-2007 at San

Ildefonso (Bulacan), Munoz-PhilRice station, and Munoz-station of Central Luzon State University

(CLSU). At all sites and in all years, tropical aerobic rice variety “Apo” (PSBRc-9) was direct dry

seeded. The treatments were the same at all sites and all years: 4 fertilizer-N levels (0, 60, 120,

and 165 kg N ha-1) and 3 irrigation treatments (two times irrigation per week (W1), one time per

week (W2), and one time in two weeks with extra irrigation from PI to flowering (W3)). At Munoz-

CLSU, irrigation treatments were slightly different: fields kept around filed capacity (W1), irrigation

applied at 75% of field capacity (W2), and irrigation applied at 50% field capacity (W3). At San

Ildefonso, the soil was clay loam (25% sand, 37% silt, and 38% clay). The groundwater depth

varied around 2 m, but during the experiments, a shallow perched water table developed that

fluctuated between 0 and 0.7 m depth. At Munoz-PhilRice, the soil was clay (46% sand, 17% silt,

and 37% clay). The groundwater depth fluctuated between 1.5 and 2 m, but during the

experiments, a shallow perched water table developed that fluctuated between 0 and 0.5 m depth.

At Munoz-CLSU, the soil was a clay loam (29% sand, 45% silt, and 26% clay). No information is

available on the depth of groundwater. At San Ildefonso and Munoz-PhilRice, irrigation water was

applied by flash-flooding, while at Munoz-CLSU, it was applied by sprinklers.

We had fertilizer N x row spacing experiments in the wet seasons (plus one experiment in

the dry season 2005) of 2004-2005 at San Ildefonso (Bulacan), Munoz-PhilRice station, and

Dapdap village (Tarlac). At all sites and in all years, tropical aerobic rice variety “Apo” (PSBRc-9)

was direct dry seeded. At all sites, the row spacings were 25, 30, and 35 cm. At San Ildefonso and


Page | 26

Dapdap, 5 fertilizer N rates were used, whereas at Munoz-PhilRice, one rate in 5 different splits

was used (Table 6.2.4). The San Ildefonso and Dapdap experiments were done in the wet seasons

of 2004 and 2005, whereas the Munoz-PhilRice experiment was done in the wet season of 2004

and the dry season of 2005. In the wet seasons, no irrigation was applied, whereas in the dry

season of 2005 at Munoz-PhilRice, crops received flush irrigation twice a week (on average). The

soil type at San Ildefonso and Munoz-PhilRice is given above (water x N experiment description).

The soil at Dapdap was a loamy sand with 78% sand, 17% silt, and 5% clay.

Table 6.2.4. N application rates and split distributions (kg ha-1) at San Ildefonso, Dapdap, and

Munoz-PhilRice in 2004 and 2005.

N-

Treatment

Basal 10-14 DAE

Early

vegetative

30-35 DAE

Mid-tillering

45-50 DAE

Panicle

Initiation

60 DAE

Flowering

Total

(kg ha-1)

San Ildefonso, Dapdap: N-amount x row spacing

N0 0 0 0 0 0 0

N60 0 18 24 18 0 60

N90 0 27 36 27 0 90

N120 0 36 48 36 0 120

N150 0 45 60 45 0 150

Munoz-PhilRice: N-split x row spacing

NS1 0 30 30 30 10 100

NS2 0 20 50 30 0 100

NS3 0 20 30 50 0 100

NS4 23 23 29 25 0 100

NS5 18 0 29 43 10 100

In Laos, we had a field experiment in 2006 on N,P,K fertilizer at two typical rainfed upland sites

(Houay Khot and Somsanouk), using 9 different rice cultivars and 3 fertilizer treatments (Table

6.2.5, Table 6.2.6). No more site information was available.


Page | 27

Table 6.2.5. Rice cultivars grown under three fertilizer regimes, Laos, 2006.

Cultivar name Type

Makhinsoung Glutinous

Nok Glutinous

Non Glutinous

Laboun Glutinous

Chao Do Non glutinous

Apo Non glutinous

B6144-MR-6-0-0 Non glutinous

Palawan Non glutinous

IR60080-46 Non glutinous

Table 6.2.6. N,P,K fertilizer application rates, Laos, 2006

Treatment no NPK Fertilizer kg ha-1

1 0, 0, 0

2 50, 0, 0

3 50, 30, 30

4 30, 30, 30

6.2.3. Results China water

First, we present the results with respect to irrigation experimentation, Table 6.2.7. The

experiments in 2001-2002 were discussed by Yang Xiaoguang et al. (2005) and Bouman et al.

(2006); the experiments in 2003-2004 are analyzed by Xue et al. (submitted, 2007), and those in

2005-2007 by Limeng Zhang et al (in prep 2008). Here, we only present major highlights. At

Changping (2001-2004), with the sandy soil and deep groundwater table, soil water tensions

varied between 10 kPa (field capacity) to beyond 90-100 kPa which was the upper limit of the

tension meters (Figures 6.2.1a-c). The highest amount of water applied (irrigation plus rainfall)

was 769 mm in W0 in 2001, and the lowest was 469 mm in W4 in 2002. At Shangzhuang (2005-

2006), the soil water tensions were much lower than at Changping because of the shallower

groundwater table that fluctuated between 0.2 and 1.2 m both years. In 2005, soil water tensions

had peak values of 30-50 kPa whereas in 2006 they stayed below 20 kPa (Figure 6.2.2). The

highest amount of water applied was 668 mm in W1 in 2005, and the lowest was 450 mm in W3 in

2006.


Page | 28

0

10

20

30

40

50

60

70

80

90

100

175 200 225 250 275 300Day number

Soil moisture tension (kPa)

PI FL H

Figure 6.2.1a. Soil water tension in treatments W1 ( ) and W3 ( ) at 15-20 cm depth

in 2001. Phenological developments are indicated PI (panicle initiation) and FL(flowering).

0

20

40

60

80

100

120

140 165 190 215 240 265 290Day number

-120

-100

-80

-60

-40

-20

0Rainfall (mm) Soil w ater potential (kPa)

FL

a

BootingPI

Tillering

Figure 6.2.1b. Soil water potential in treatments W1 ( ), W2 ( ) and W3 ( ) at

15-20 cm depth in 2003. Daily rainfall is shown by black solid bars. Phenological developments are

indicated by arrows: tillering, PI (panicle initiation), booting, FL(flowering).


Page | 29

0

20

40

60

80

100

120

140 165 190 215 240 265 290

Day number

-120

-100

-80

-60

-40

-20

0Rainfall (mm) Soil w ater potential (kPa)

PI FL

b

Tillering Booting

Figure 6.2.1c. Soil water potential in treatments W1 ( ), W2 ( ) and W3 ( ) at

15-20 cm depth and 2004. Daily rainfall is shown by black solid bars. Phenological developments

are indicated by arrows: tillering, PI (panicle initiation), booting, FL(flowering).

Figure 6.2.2. Soil water tension in treatments W1, W2, and w3 in 2005 (left panel) and 2006 (right

panel) versus days after emergence.

0

10

20

30

40

50

60

0 20 40 60 80 100 120 140Days after emergence

W1

W2

W3

0

10

20

30

40

50

60

0 20 40 60 80 100 120 140

Days after emergence

W1

W2

W3


Page | 30

Table 6.2.7. Amounts of water inputs by irrigation plus rainfall (mm) per water treatments at

Changping, 2001-2004 and Shangzhuang, 2005-2006.

Treatment 2001 2002 2003 2004 2005 2006

W0 644 769

W1 577 708 688 705 668 550

W2 586 620 618 675 526 490

W3 519 695 648 645 484 450

W4 469 547 578 605

Our yield levels (Table 6.2.8) compare well with other reports for aerobic rice HD297 in northern

China. At another site close to Beijing, Tao et al. (2006) reported yields of 5.7-6.1 t ha-1 with

irrigation applied when the soil water tension at 15 cm depth exceeded 15 kPa, resulting in around

1400 mm total water input. In field experiments near Kaifeng (34˚82΄N, 114˚51΄E; 69 m asl),

Feng et al. (2007) obtained relatively lower yields of 2.4-3.6 t ha-1 with 750-1000 mm total water

input. In the same area, however, Bouman et al. (2007) reported farmers’ yields of HD297 up to

5.5 t ha-1 with sometimes even as little as 566 mm total water input but with groundwater depth

varying between 20 and 200 cm.

In our experiments, highest yields of 5.3-5.6 t ha-1 were obtained at Changping, despite

the drier soil conditions there, while the yields at Shangzhuang (with wetter soil conditions) were

not higher than 5.1 t ha-1 (Table 6.2.8). Even with low amounts of total water input, yields were

still higher than 2.5 t ha-1 in all years except in 2003 where yield declined to 0.5 t ha-1 in W4. The

low yields in 2003 were mainly caused by very low grain filling realized in all treatments, but

especially so in W3 and W4 (Table 6.2.9). Around the time of flowering, from 77 to 97 days after

emergence (DAE), less than 10 mm rainfall was recorded, with only 40 mm irrigation applied in

W1 and W2, and no irrigation in W3 and W4. As a consequence, soil water tensions went up to

above 100 kPa (data outside tensiometer measurement range). Rice is very sensitive to drought

around flowering which leads to increased spikelet sterility and decreased grain filling (Cruz and

O’Toole, 1984; Ekanayake et al., 1989). In the other years, soil water tensions at flowering were

not as high as in 2003, and percentages of grain filling and yields were higher (Bouman et al.,

2006).


Page | 31

Table 6.2.8. Yield of HD297 (averaged over N levels) (t ha-1) for 5 water treatments at Changping,

2001-2004 and Shangzhuang, 2005-2006.

Treatment 2001 2002 2003 2004 2005 2006

W0 4.7 5.3

W1 4.3 4.7 4.4 5.6 5.1 4.4

W2 4.2 3.9 3.4 5.4 4.7 4.3

W3 3.4 4.6 1.4 5.4 4.7 4.1

W4 2.5 3.0 0.5 5.0

Table 6.2.9. Percentage grain filling (%) in the 2002-2004 field experiments at Changping, per

irrigation treatment and fertilizer-N treatment.

Water treatment

Year N amount

(kg ha-1) W0 W1 W2 W3 W4

2002 none 65 67 58 66 52

2003 0 70 68 30 12

113 60 62 26

150 57 64 26

225/3splits 52 56 20 8

225/5splits 53 59 12 10

2004 0 84 86 85 83

75 83 82 82 81

125 83 80 83 79

175 81 77 84 80

225 82 83 80 80

The trends in water productivity followed mainly the trends in yield since differences among

amounts of water inputs were smaller than differences among yield (Table 6.2.10). Highest values

were 0.88-0.95 kg grain m-3 water across both sites, whereas the lowest values were obtained at

Changping in 2003 with the low percentages grain filling (see above). The water productivities are


Page | 32

higher than the most common values of 0.3-0.4 kg grain m-3 water obtained in lowland rice

(Tuong et al., 2005).

Table 6.2.10. Water productivity with respect to irrigation plus rainfall of HD297 at the highest N

level (WPIR, kg grain m-3 water) for 5 water treatments at Changping, 2001-2004 and

Shangzhuang, 2005-2006.

Treatment 2001 2002 2003 2004 2005 2006

W0 0.73 0.58

W1 0.75 0.61 0.70 0.84 0.75 0.66

W2 0.71 0.63 0.60 0.84 0.88 0.72

W3 0.65 0.67 0.25 0.88 0.95 0.73

W4 0.54 0.54 0.10 0.88

On average across 2001-2004 at Changping, the computed evapotranspiration in the ‘wettest’

irrigation treatment was 596 mm (Table 6.2.11; which is 5.2 mm d-1). Using the average amounts

of rainfall in three typical years (Table 6.2.12), the calculated net irrigation requirements are listed

in Table 6.2.13.

Table 6.2.11. Average calculated evapotranspiration (ET, mm) per growth stage, Changping 2001-

2004.

Growth Stage Evapotranspiration

(mm)

Emergence-PI 286

PI-Booting 64

Booting-Flowering 91

Flowering-Maturity 155

Whole season 596


Page | 33

Table 6.2.12. Average rainfall (mm) per growth stage for three types of years, calculated from 50

years of weather data at Changping.

Sowing-

Emergence

Emergence-

PI PI-Booting

Booting-

Flowering

Flowering-

Maturity

Whole

season

Wet

(25%) 10 315 190 50 95 625

Average

(50%) 4 245 130 30 85 510

Dry

(75%) 0 145 85 25 45 335

Table 6.2.13. Net irrigation requirements calculated as evapotranspiration minus rainfall(mm) per

growth stage for three types of years, calculated from 50 years of weather data at Changping.

Sowing-

Emergence

Emergence-

PI PI-Booting

Booting-

Flowering

Flowering-

Maturity

Whole

season

Wet

(25%) 66 0 0 39 62 167

Average

(50%) 72 44 0 58 72 246

Dry

(75%) 76 144 0 63 112 395

6.2.4. Results China water x nitrogen

Final aboveground biomass and grain yield for all experiments with water x nitrogen interaction

are given in Table 6.2.14 for Changping (2003-2004), and in Table 6.2.15 for Shanzhuang (2005-

2006). At Changping, irrigation had a significant effect on yield in all three experiments. Yields

were highest in W1 and consistently declined through W2 to W3 and W4, though not all differences

among treatments were significant. The application of any amount of fertilizer N either reduced

yield and biomass (2003) or kept yield and biomass at the same level as without N fertilization

(2004). In experiment 1, the application of N decreased yield but the difference with 0 N was not

significant. In experiment 2, the application of N also decreased yield, but the difference between

0 N and 225 kg N ha-1 was significant only with three splits of the fertilizer. The interaction

between irrigation and N was significant in experiment 2 but not in experiment 1. Like with


Page | 34

biomass, splitting the fertilizer N in three gave significantly lower yield in W1 (and a bit in W4) but

higher yield in W2 and W3 than splitting it in five. In experiment 3, the differences among N levels

were small and inconsistent. Although the highest yield was obtained with 75 kg N ha-1, significant

difference only occurred between 75 and 225 kg N ha-1 levels.


Page | 35

Table 6.2.14. Total aboveground biomass and grain yield of aerobic rice for different irrigation and

N treatments at Changping, Beijing, 2003-2004

Biomass (t ha-1) Grain yield (t ha-1) Experiment

N rate

(kg ha-1) W1 W2 W3 W4 Mean W1 W2 W3 W4 Mean

0 17.4 18.0 15.7 - 17.0 4.46 4.06 2.38 - 3.63

113 18.3 16.4 14.0 - 16.2 3.91 3.68 2.50 - 3.36

150 16.2 13.7 14.3 - 14.7 4.39 3.35 1.44 - 3.06

Mean 17.3 16.0 14.7 - 16.0 4.25 3.70 2.11 - 3.35

W 1.54a 1.00

N 2.26 N/S

2003, Exp1

W*N N/Sb N/S

0 14.7 11.2 16.2 13.2 13.8 3.27 2.60 1.58 0.60 2.01

225/3splits 13.8 12.3 11.7 11.5 12.3 2.44 2.67 1.42 0.25 1.70

225/5splits 15.8 11.9 10.6 13.3 12.9 3.61 2.36 0.77 0.53 1.82

Mean 14.8 11.8 12.8 12.7 13.0 3.11 2.54 1.26 0.46 1.84

W 1.27 0.48

N 1.11 0.24

2003, Exp2

W*N 2.05 0.58

0 14.7 14.6 14.4 14.7 14.6 5.58 5.37 5.38 5.07 5.35

75 16.5 17.8 17.3 14.9 16.6 6.03 5.70 5.29 5.50 5.63

125 17.7 12.0 19.1 14.9 15.9 5.37 5.34 5.64 5.04 5.35

175 17.0 17.2 15.7 13.6 15.9 5.41 4.78 5.39 4.73 5.08

225 16.5 17.5 14.3 18.2 16.6 5.53 5.57 5.06 4.61 5.19

Mean 16.5 15.8 16.2 15.2 15.9 5.58 5.35 5.35 4.99 5.32

W N/S 0.25

N N/S 0.39

2004, Exp3

W*N N/S N/S

a LSD at P = 0.05; N/S means there was no significant difference at P = 0.05

At Shangzhuang, the only significant effect was by water (no significant effects of nitrogen, year,

or interaction of year x water x nitrogen), Table 6.2.15. Yields were higher in W1 than in W3 in all

cases. The addition of fertilizer-N reduced yields from the 0-N controls in 3 out of the 4 W x N


Page | 36

combinations. In 2005, we noticed severe lodging of the crop at flowering, and the degree of

lodging seemed to increase with increasing N application rate.

Table 6.2.15. Total aboveground biomass and grain yield of aerobic rice for different irrigation and

N treatments at Shangzhuang, Beijing, 2005-2006.

Year Irrigation

treatment

Fertilizer Nitrogen

treatment

Biomass (t ha-1) Yield (t ha-1)

2005 W1 N0 (0 kg N ha-1) 15.0 5.3

N2 (150 kg N ha-1) 14.0 4.5

W3 N0 11.5 5.0

N2 12.1 4.3

2006 W1 N0 9.3 5.0

N2 10.7 4.7

W3 N0 10.1 4.6

N2 9.8 4.3

Year ** Ns

Water Ns *

Nitrogen Ns Ns

Water x nitrogen Ns Ns

Year x water x nitrogen Ns Ns

Source of variation

Pr>F; <0.001:***; <0.01:**; <0.05:*; n.s.: not signification.

The lack of positive response of yield and crop growth parameters to N fertilization in all our

experiments (coupled with a “luxurious N uptake”, data not shown here, see Xue et al.,

submitted), indicates a high level of indigenous N supply (native soil N supply, N in irrigation

water, atmospheric N deposition). A high level of soil N supply may have been the result of

continuous over fertilization of the maize and wheat crops that preceded our experiments for the

last 5-10 years. Around Beijing, the average N application rates in farmers’ fields are 309 kg N ha-1

in winter wheat and 256 kg N ha-1 in maize (Zhao et al., 1997). Ma (1999) reported average

combined N application rates of more than 500 kg N ha-1 in wheat-maize cropping systems in

nearby Shandong Province. Such high levels of N application have resulted in increased soil N

contents (Liu et al., 2003). Additionally, about 40 kg atmospheric N ha-1 is deposition annually in

the Beijing area which contributes to indigenous soil N (Liu et al., 2006).


Page | 37

6.2.5. Results China NPK

Rainfall in the aerobic rice season was 1098 mm in 2005 and 629 mm in 2006. The irrigation

inputs were estimated as 180 mm in 2005, and 240 mm in 2006. Because of heavy rainfall in

2005, the groundwater level increased from around 2 m depth to 0-0.8 m during the whole aerobic

rice crop in 2005 (Figure 6.2.3). Consequently, the soil was very wet and the soil water tension at

15-20 cm depth remained within 0-15 kPa (Figure 6.2.4). In 2006, with less rainfall, the

groundwater fluctuated around 0.6 m depth during July-August, but was about 1.8 m for the rest

of the season. From mid August onward, the soil water tension ranged between 20 and 100 kPa.

Figure 6.2.3. Groundwater depth in during the aerobic rice crop, at Shuanghu, 2005 and 2006.

Figure 6.2.4. Soil moisture tension at 15-20-cm depth during the aerobic rice crop, at Shuanghu,

2005 and 2006.


Page | 38

Fertilizer applications significantly affected the aboveground biomass and grain yields of aerobic

rice and winter wheat (Table 6.2.16). Aboveground biomass of aerobic rice ranged from 7.0 to 9.7

t ha-1 in 2005 and from 6.3 to 9.0 t ha-1 in 2006. For winter wheat, biomass yield ranged from 5.4

to 13.6 t ha-1 in AR-WW, and from 10.0 to 14.9 t ha-1 in WW-AR. The reduction in biomass was

mainly the consequence of N omission. A cumulative N omission in winter wheat following aerobic

rice caused a more severe reduction in biomass (6-7 t ha-1) than in winter wheat preceding aerobic

rice (4-5 t ha-1).

Yields of aerobic rice varied from 3.2 to 4.1 t ha-1 in 2005 and from 2.9 to 4.0 t ha-1 in 2006, and

of winter wheat from 2.7 to 7.1 t ha-1 in AR-WW and from 4.3 to 6.6 t ha-1 in WW-AR. Again, the

yield loss due to cumulative N omission was much bigger in the wheat crop following aerobic rice.

The losses in yield due to N omission were larger in winter wheat than in aerobic rice, indicating

that the N demand of wheat was higher than that of aerobic rice, and that it could not be met from

indigenous N supply.

The aboveground biomass and yield of aerobic rice and winter wheat were significantly

reduced in the PK treatment (N omission) compared with the full NPK treatment. The reduction in

grain yield was 0.5 (14%) and 0.8 t ha-1 (21%) for aerobic rice, in 2005 and 2006 respectively,

and 2.3 (35%) and 4.3 t ha-1 (61%) for winter wheat in the subsequent years. There was no

significant difference in either aboveground biomass or yield among NPK, NK (P omission) and NP

(K omission) treatments in aerobic rice. P and K omissions caused significant reductions in yield of

wheat in the AR-WW sequence: in the P and K omission treatments, the yield loss was 1.6 (23%)

and 1.2 t ha-1 (17%).


Page | 39

Table 6.2.16. Aboveground biomass and grain yield of aerobic rice (AR) and winter wheat (WW)

for two AR-WW (June 2005 to June 2006) and WW-AR (October 2005 to October 2006) cropping

sequences.

Treatment Aerobic rice Winter wheat

Biomass (t ha-1) Yield (t ha-1) Biomass (t ha-1) Yield (t ha-1)

AR-WW

NPK 9.7 a 3.7 a 13.6 a 7.1 a

PK 7.3 b 3.2 b 6.7 c 2.8 c

NK 9.6 a 4.1 a 11.5 b 5.5 b

NP 9.6 a 4.1 a 12.7 ab 5.9 b

CK 7.0 b 3.3 b 5.4 c 2.7 c

WW-AR

NPK 9.0 a 3.8 a 14.9 a 6.6 a

PK 6.5 b 3.0 b 10.3 bc 4.3 b

NK 8.1 a 3.7 a 13.1 ab 6.1 a

NP 8.8 a 4.0 a 13.7 a 6.4 a

CK 6.3 b 2.9 b 10.0 c 4.4 b

For each cropping sequence, different letters within a column indicate significant differences at

P<0.05.

6.2.6. Results India water

At the WTC-IARI station in 2005, the amount of irrigation plus effective rainfall was 1314 mm in

I0, 1014 mm in I20, and 894 mm in I40. In 2006, these amounts were 100 mm in I0, 980 mm in

I20, and 780 mm in I40. Compared with typical amounts of water applied to lowland rice fields

(under alternate wetting and drying), these amounts translate into water savings of the order of

30-40% for production levels of 4 t ha-1. The soil water tensions in all three irrigation treatments

slightly exceeded the intended target on a few occasions but were mostly within the intended

ranges (Figure 6.2.5). Yields were highest in I0 and significantly decreased for most varieties with

decreasing amount of irrigation water in I20 and I40 (Figures 6.2.6a,b). Water productivity with

respect to irrigation plus rainfall of Pusa Rice Hybrid 10, Pusa 834, IR74371-46-1-1, IR55423-01

(Apo1) and Proagro 6111 was about 0.42-0.47 kg m-3 at I20 irrigation, while it was 0.50 kg m-3 at

I40 irrigation (Figure 6.2.7).


Page | 40

Figure 6.2.5. Soil water tension (kPa) measured at different depths in treatments I0, I20, and I40

in the variety screening trial at WTC-IARI station, Delhi, 2005.

0

10

20

30

40

50

60

70

80

40 60 80 100 120 140

Days after sowing

Soi

l moi

stur

e te

nsio

n (k

Pa)

.

10 cm 20 cm 30 cm 40 cm

0

10

20

30

40

50

60

70

80

40 60 80 100 120 140

Days after sowing

Soi

l moi

stur

e te

nsio

n (k

Pa)

.

10 cm 20 cm 30 cm 40 cm

0

10

20

30

40

50

60

70

80

40 60 80 100 120 140

Days after sowing

Soi

l moi

stur

e te

nsio

n (k

Pa)

.

10 cm 20 cm 30 cm 40 cm

I I2

I4


Page | 41

Figure 6.2.6a. Yield of different genotypes under aerobic soil conditions at three irrigation levels at

the WTC-IARI station, Delhi, 2005.

0

100

200

300

400

500

600

Proagr

o 6111

Pusa 83

4

Pusa 67

7

Pusa 14

62-03

-14-1

Pusa R

H 10

IR55

419-0

4

IR55

423-0

1

IR71

700-2

47-1-

1-2

IR72

875-9

4-3-3-

2

IR74

371-5

4-1-1

Pusa Sugan

dh 4

IR74

371-4

6-1-1

Mea

n

Gra

in y

ield

(g

m-2

)

I0 I20 I40


Page | 42

Figure 6.2.6b. Yield of different genotypes under aerobic soil conditions at three irrigation levels at

the WTC-IARI station, Delhi, 2006.

Figure 6.2.7. Water productivity with respect to irrigation plus effective rainfall (kg m-3) of different

genotypes under aerobic soil conditions at three irrigation levels at the WTC-IARI station, Delhi,

2006.

0

100

200

300

400

500

600

Pro

agro

611

1

Pus

a 83

4

Pus

a 67

7

Pus

a 14

62-0

3-14

-1

Pus

a R

H10

IR55

419-

04

IR55

423-

01

IR71

700-

247-

1-1-

2

IR72

875-

94-3

-3-2

IR74

371-

54-1

-1

Pus

a S

ugan

dh 4

IR74

371-

46-1

-1

Inda

m 1

00 0

01

Raj

alax

mi

Pus

a S

ugan

dh 5

Anj

ali

Kal

inga

Pus

a S

ugan

dh 3

Nav

een

Ann

ada

Van

dana

Sar

bhat

i

IM10

9

N22

Pad

dy Y

ield

(g

m-2)

I0 I20 I40

0.00

0.10

0.20

0.30

0.40

0.50

0.60

Pro

agro

611

1

Pus

a 83

4

Pus

a 67

7

Pus

a 14

62-0

3-14

-1

Pus

a R

H10

IR55

419-

04

IR55

423-

01

IR71

700-

247-

1-1-

2

IR72

875-

94-3

-3-2

IR74

371-

54-1

-1

Pus

a S

ugan

dh 4

IR74

371-

46-1

-1

Inda

m 1

00 0

01

Raj

alax

mi

Pus

a S

ugan

dh 5

Anj

ali

Kal

inga

Pus

a S

ugan

dh 3

Nav

een

Ann

ada

Van

dana

Sar

bhat

i

IM10

9

N22

Wat

er P

rodu

ctiv

ity (

kg m

-3)

I0 I20 I40


Page | 43

6.2.7. Results Philippines water x nitrogen

The amount of water put in by irrigation in W1 was higher in the Philippines than in China (Table

6.2.17). Using surface irrigation, the lowest amount of irrigation water used was 480 mm in San

Ildefonso, and the highest was 1070 mm in Munoz-PhilRice (with 22 to 200 mm rainfall). Using

sprinkler irrigation in Munoz-CLSU, the amount of combined irrigation and rain water varied from

as little as 274 mm to as much as 1300 mm.

Table 6.2.17. Irrigation water input (mm) per water treatment, rainfall during the experiment

(mm), and fertilizer-N inputs per nitrogen treatment (kg ha-1), at Munoz (PhilRice) and san

Ildefonso (BASC/BSWM), 2005-2007 dry seasons.

Site Year W1 W2 W3 Rain N1 N2 N3 N4

Munoz

(PhilRice)

2005

1070 630 590

22 0 60 120 165

Munoz

(PhilRice)

2006

1040 680 560

200 0 60 120 165

Munoz

(CLSU)

2006 681 418 378 Included 0 60 120 165

Munoz

(CLSU)

2007 1300 280 274 Included 0 60 120 165

San

Ildefonso

2006 900 540 480 105 0 60 120 165

San

Ildefonso

2007 970 610 510 230 0 60 120 165

The soil was much wetter in the Philippine experiments than in the China experiments, as

evidenced by the much lower soil moisture tensions (Figures 6.2.8a-d). Despite that the fields

were never flooded, the soil water tensions stayed mostly within 0-10 kPa for all treatments at

Munoz-PhilRice in 2005, within 0-20 kPa at Munoz-PhilRice in 2006, within 0-20 kPa at San

Ildefonso in 2006, and within 0-30 kPa at San Ildefonso in 2007 (no tensiometer data at Munoz-

CLSU). The lack of clear differences in soil water tension among irrigation treatments, and the

relatively low levels of soil water tension were caused by the development of perched groundwater

tables underneath the experiments (Figures 6.2.9a-c). At Munoz-PhilRice in 2005, the groundwater

depth as measured tubes installed in bunds surrounding the experiment fluctuated between 1.5

and 2 m, but the groundwater measured in tubes installed within the experimental fields fluctuated


Page | 44

between 0 and 0.5 m depth. At San Ildefonso, in 2007, the “real” groundwater depth varied

around 2 m, but underneath the experimental field, it fluctuated between 0 and 0.7 m depth.

Figure 6.2.8a. Soil water tension (kPa) at 15-35 cm depth in water x N trial, Munoz (PhilRice), dry

season 2005.

0.0

5.0

10.0

15.0

20.0

25.0

08-Jan 28-Jan 17-Feb 09-Mar 29-Mar 18-Apr 08-May

W1

W2

W3

Figure 6.2.8b. Soil water tension (kPa) at 15-35 cm depth in water x N trial, Munoz (PhilRice), dry

season 2006.

0.0

5.0

10.0

15.0

20.0

25.0

30.0

23-Jan 12-Feb 4-Mar 24-Mar 13-Apr 3-May 23-May

W1

W2

W3


Page | 45

Figure 6.2.8c. Soil water tension (kPa) at 15-35 cm depth in water x N trial, San Ildefonso, dry

season 2006.

0.0

5.0

10.0

15.0

20.0

25.0

2006

/02/

18

2006

/02/

25

2006

/03/

04

2006

/03/

11

2006

/03/

18

2006

/03/

25

2006

/04/

01

2006

/04/

08

2006

/04/

15

2006

/04/

22

2006

/04/

29

W1

W2

W3

Figure 6.2.8d. Soil water tension (kPa) at 15-35 cm depth in water x N trial, San Ildefonso, dry

season 2007.

0

5

10

15

20

25

30

35

40

1-M

ar

8-M

ar

15-M

ar

22-M

ar

29-M

ar

5-A

pr

12-A

pr

19-A

pr

26-A

pr

3-M

ay

10-M

ay

17-M

ay

24-M

ay

W1W2W3


Page | 46

Figure 6.2.9a. Groundwater depth (cm) measured at three locations in water x N trial, Munoz

(PhilRice), dry season 2005.

-300

-250

-200

-150

-100

-50

0

50

1-Ja

n

8-Ja

n

15-J

an

22-J

an

29-J

an

5-F

eb

12-F

eb

19-F

eb

26-F

eb

5-M

ar

12-M

ar

19-M

ar

26-M

ar

2-A

pr

9-A

pr

16-A

pr

23-A

pr

30-A

pr

Groundwater table depths (cm)

Figure 6.2.9b. Perched groundwater table depth (cm) measured at three locations in water x N

trial, Munoz (PhilRice), dry season 2005.

-80-70-60-50-40-30-20-10

010

1-Ja

n

8-Ja

n

15-J

an

22-J

an

29-J

an

5-F

eb

12-F

eb

19-F

eb

26-F

eb

5-M

ar

12-M

ar

19-M

ar

26-M

ar

2-A

pr

9-A

pr

16-A

pr

23-A

pr

30-A

pr

Perched table depths (cm)

Figure 6.2.9c. Perched groundwater table depth in water x N trial, San Ildefonso, dry season 2007.

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

10

21-F

eb

28-F

eb

7-M

ar

14-M

ar

21-M

ar

28-M

ar

4-A

pr

11-A

pr

18-A

pr

25-A

pr

2-M

ay

9-M

ay

16-M

ay

23-M

ay

30-M

ay

cm Mean Groundwater level


Page | 47

Highest yields were 5.4 t ha-1 in Munoz-PhilRice 2005, 2.9 t ha-1 in Munoz-PhilRice 2006, 3.4 t ha-1

in Munoz-CLSU 2006, 3.8 t ha-1 in Munoz-CLSU 2007, 3.1 t ha-1 in San Ildefonso 2006, and 3.5 t

ha-1 in San Ildefonso 2007 (Tables 6.2.18a-f). The yields were a bit lower in Munoz-CLSU with

sprinkler irrigation. Because of the absence of differences in soil water tensions among treatments,

there were no significant effects of irrigation water regime in most experiments. Only in Munoz-

CLSU was yield much higher in W1 than in W2 and W3, in both 2006 and 2007, though across all

water treatments, the effect of water was not significant. The effect of N was significant at the 1%

or 5% level in all experiments: yields increased with increasing N application rate, especially in

Munoz (both sites). At San Ildefonso, the yield increase with added N was smaller than at Munoz.

Table 6.2.18a. Yield (t ha-1) of Apo in water x N trial, Munoz (PhilRice), dry season 2005.

Mean W1 W2 W3 Ave

N1 1.39 1.24 1.41 1.35

N2 3.15 2.73 2.54 2.81

N3 4.00 4.50 4.54 4.34

N4 5.02 5.43 5.25 5.23

Ave 3.39 3.47 3.43

Effect of N is significant at the 1% level; effect of water is not significant.

Table 6.2.18b. Yield (t ha-1) of Apo in water x N trial, Munoz (PhilRice), dry season 2006.

Mean W1 W2 W3 Ave

N1 0.78 0.90 0.88 0.85

N2 1.47 1.30 1.14 1.30

N3 2.61 2.61 2.31 2.51

N4 2.92 2.68 2.26 2.62

Ave 1.94 1.87 1.65



Page | 48

Table 6.2.18c. Yield (t ha-1) of Apo in water x N trial, Munoz (CLSU), dry season 2006.

Mean W1 W2 W3 Ave

N1 1.67 1.89 1.27 1.61

N2 2.26 2.02 1.67 1.98

N3 2.60 1.63 2.31 2.18

N4 3.43 2.04 2.25 2.57

Ave 2.49 1.89 1.88

Effect of N is significant at the 5% level; effect of water is not significant

Table 6.2.18d. Yield (t ha-1) of Apo in water x N trial, Munoz (CLSU), dry season 2007.

Mean W1 W2 W3 Ave

N1 2.25 1.98 1.24 1.82

N2 3.52 2.08 1.79 2.46

N3 3.77 2.21 2.01 2.66

N4 3.56 2.16 1.83 2.52

Ave 3.28 2.10 1.72

Effect of N and water is significant at the 5% level.

Table 6.2.18e. Yield (t ha-1) of Apo in water x N trial, San Ildefonso, dry season 2006.

Mean W1 W2 W3 Ave

N1 2.41 2.11 2.48 2.33

N2 2.31 2.65 2.58 2.51

N3 2.97 2.79 2.69 2.82

N4 2.77 3.03 3.09 2.96

Ave 2.61 2.65 2.71



Page | 49

Table 6.2.18f. Yield (t ha-1) of Apo in water x N trial, San Ildefonso, dry season 2007.

Mean W1 W2 W3 Ave

N1 2.34 2.53 3.05 2.64

N2 3.25 3.61 3.25 3.37

N3 3.38 3.45 3.05 3.30

N4 3.78 2.99 3.54 3.44

Ave 3.19 3.15 3.22

Effect of N and water are significant at the 1% level.

6.2.8. Results Philippines nitrogen x row spacing

Here, we present the results only with respect to N application rate and distribution (there is no

effect of row spacing: see paragraph “Crop establishment” for analysis). More detailed results are

presented by Lampayan et al. (in prep). Total seasonal rainfall was 1210 mm in 2004 and 911 mm

in 2005 at San Ildefonso, 984 mm in 2004 and 897 mm in 2005 at Dapdap, and 1492 mm in 244

and 20 mm in 2005 (dry season!) at Munoz-PhilRice. Groundwater depths are given in Figure

6.2.10. Groundwater tables were relatively deep at Munoz-PhilRice, at 1-1.8 m. Groundwater

tables were relatively shallow in 2004 in San Ildefonso and Dapdap where they reached into the

root zone. Except for Dapdap in 2005, soil water tensions measured in the root zone usually varied

between 0 and 15 kPa only (Figure 6.2.11). At Dapdap in 2005, soil water tensions went up to 35-

40 kPa.


Page | 50

Figure 6.2.10. Groundwater depth (cm) in the N x row spacing experiments in San Ildefonso

(Bulacan), Dapdap, and Munoz-PhilRice, in 2004-2005.

(a) Bulacan

(b) Dapdap

( c) PhilRice

WS2004

-200-180-160-140-120-100

-80-60-40-20

020

150 180 210 240 270

Gro

undw

ater

dep

th (

cm)

Day number

WS2005

150 180 210 240 270

Day number

WS2004

-200-180-160-140-120-100

-80-60-40-20

020

150 180 210 240 270

Gro

undw

ater

dep

th (

cm)

Day number

WS2005

150 180 210 240 270

Day number

WS2004

-200-180-160-140-120-100

-80-60-40-20

020

150 180 210 240 270

Gro

undw

ater

dep

th (

cm)

Day number

DS2005

0 30 60 90 120Day number


Page | 51

Figure 6.2.11. Soil water tension (kPa) at 10-15 cm depth in the N x row spacing experiments in

San Ildefonso (Bulacan), Dapdap, and Munoz-PhilRice, in 2004-2005.

(a) San Ildefonso

(b) Dapdap

( c) PhilRice

0

5

10

15

20

25

30

35

40

150 180 210 240 270 300

WS2004

Soi

l ten

sion

, KP

a

Day number150 180 210 240 270 300

WS2005

Day number

05

10152025303540

150 180 210 240 270 300

WS2004

Soi

l ten

sion

, KP

a

Day number150 180 210 240 270 300

WS2005

Day number

05

10152025303540

150 180 210 240 270 300

WS2004

Soi

l ten

sion

, KP

a

Day number

05

10152025303540

0 30 60 90 120 150

DS2005

Day number


Page | 52

In the N rate experiment, maximum yields in the different sites and years were 4-4.5 t ha-1 (Table

6.2.19). Fertilizer N addition had a significant positive effect on yield. At San Ildefonso, highest

yields were obtained at 60-120 kg N ha-1 (statistically the same), and in Dapdap at 120-150 kg N

ha-1 (statistically the same). Except in San Ildefonso in 2005, there was considerable lodging in the

120 and 150 kg N ha-1 treatments (Table 6.2.20), and we can speculate that yields would have

been higher without lodging.

Table 6.2.19. Yeld and total biomass (averaged over row spacings) of Apo in different N application

treatments at San Ildefonso and Dapdap, 2004 and 2005 wet seasons.

Site and year

N treatment

Yield (t ha-1) Total biomass

at harvest (t ha-1)

San Ildefonso, 2004 N0 2.18 a 5.52 a

N60 4.13 b 12.94 b

N90 4.64 b 16.22 c

N120 4.04 b 15.23 c

N150 3.62 b 15.11 c

San Ildefonso, 2005 N0 2.52 a 6.70 a

N60 3.77 b 9.79 b

N90 4.11 c 10.35 bc

N120 4.28 c 10.22 bc

N150 4.28 c 11.48 c

Dapdap, 2004 N0 1.69 a 4.65 a

N60 3.31 b 9.28 b

N90 3.98 c 11.19 c

N120 4.48 d 11.84 cd

N150 4.85 d 12.46 d

Dapdap, 2005 N0 1.67 a 3.86 a

N60 3.07 b 6.93 b

N90 3.35 bc 6.88 b

N120 3.77 cd 8.42 c

N150 4.00 d 9.27 c

Means followed by a different letter (within columns) are significantly different at the 5% level.


Page | 53

Table 6.2.20. Degree of lodging (%) of Apo at San Ildefonso and Dapdap, 2004 and 2005 wet

seasons.

Site and year N treatment

San Ildefonso

WS2004

San Ildefonso

WS2005

Dapdap WS2004 Dapdap WS2005

N0 0 a 0 a 0 a 0 a

N60 3 a 0 a 0 a 2 a

N90 42 b 0 a 8 a 4 a

N120 56 bc 0 a 46 b 44 b

N150 62 cd 0 a 76 c 67 b


In the N splitting experiment, yields were 2-3.9 t ha-1 in the wet season and 5.4-6.1 t ha-1 in the

dry season (Table 6.2.21). There was hardly any significant effect of N splitting on yield, except

that the treatments with a late application of 10 kg N ha-1 at flowering (NS1 and NS5) produced

significantly higher yields in 3 out of the 4 cases.


Page | 54

Table 6.2.21. Grain yield and total biomass (averaged over row spacings) of Apo in different N split

treatments at Munoz-PhilRice, 2004 wet season and 2005 dry season. N treatments are explained

in the Method section above.

Year

N treatment


at harvest (t ha-1)

2004 NS1 3.90 b 10.0 a

NS2 3.48 a 9.6 a

NS3 3.46 a 9.5 a

NS4 3.24 a 9.6 a

NS5 3.36 a 9.1 a

2005 NS1 6.10 b 12.1 b

NS2 5.80 ab 11.6 ab

NS3 5.69 ab 11.6 ab

NS4 5.40 a 10.8 a

NS5 6.10 b 11.6 ab


6.2.9. Results Laos NPK

Results of the variety x nutrient trials are given in Table 6.2.22 and Table 6.2.23. Significant

differences in grain yield were observed among rice cultivars at the two sites. No significant

differences in grain yield were observed among fertilizer treatments at the two sites. However,

yields were consistently highest in the 50-30-30 NPK treatment for all cultivars at Houy Khot,

while they were consistently highest in the 50-0-0 NPK treatment at Somsanouk. Finally, grain

yield was, on the average, higher at Somsanouk (fallowed for four years) than at Houy Khot

(continuously cropped to upland rice for the last three years).


Page | 55

Table 6.2.22. Grain yield (t ha-1) for different cultivars under 4 nutrient treatments, Houay Khot,

2006.

Fertilizer treatment kg (N-P-K) ha-1

Variety 0-0-0 50-0-0 50-30-30 30-30-30

Makhinsoung 0.97 1.38 1.91 1.04

Nok 0.84 1.21 1.86 1.55

Non 2.18 2.06 2.65 2.21

Laboun 1.66 1.72 2.00 1.83

Chaodor 1.27 1.09 1.34 1.24

Apo 2.50 2.57 3.11 2.48

B6144-MR-6-0-0 2.14 2.42 3.11 2.71

Palawan 0.94 1.03 1.54 1.40

IR60080-6A 1.87 2.07 3.19 2.67

Table 6.2.23. Grain yield (t ha-1) for different cultivars under 4 nutrient treatments, Somsanouk,

2006.

Fertilizer treatment kg (N-P-K) / ha

Variety 0-0-0 50-0-0 50-30-30 30-30-30

Makhinsoung 3.49 4.65 4.31 3.83

Nok 3.72 3.95 3.82 4.02

Non 4.42 4.01 4.00 4.08

Laboun 3.39 3.99 3.96 3.88

Chaodor 3.02 3.43 3.21 2.71

Apo 5.07 4.75 3.95 4.90

B6144-MR-6-0-0 3.95 4.44 3.84 4.13

Palawan 3.94 3.94 3.31 3.94

IR60080-6A 3.94 4.15 3.78 3.94


Page | 56

6.2.10. Conclusion

China. In Beijing, the highest aerobic rice yields of HD297 were 5-5.6 t ha-1 with 600-700 mm

total irrigation plus rainfall water (and in Shangzhuang 2005-2007, with any unaccounted capillary

rise). Yields were 2.5-4.3 t ha-1 with 450-550 mm water input. Very low yields (down to 0.5 t ha-

1) were obtained with dry soil (water tension higher than 100 kPa) around flowering. The main

reason for the low yields was the high spikelet sterility. Irrigation is essential at flowering time if

there is no rain. The average seasonal evapotranspiration (ET) requirement was about 600 mm.

With shallow groundwater, capillary rise can meet most of the ET and there may be no need to

irrigate. With deep groundwater tables, the net irrigation needs are 167 mm in a typical ‘wet

rainfall year” (2 times irrigation), 246 mm in a typical “average rainfall year” (3 times irrigation),

and 395 mm in a typical “dry rainfall year” (4-5 times)

In Beijing, the application of any amount of fertilizer N either reduced yield and biomass

(2003, 2005, 2006 experiments) or kept yield and biomass at the same level as without N

fertilization (2004 experiments). Fields in northern China cropped to summer maize and/or winter

wheat may have been over fertilized for many years to the extent that they are now ‘saturated’

with N. Moreover, they may receive large amounts of N through atmospheric deposition (around

40 kg atmospheric N ha-1 in the Beijing area). To assist farmers with fertilizer N management in

aerobic rice, an inventory is needed of indigenous soil N supply in the target area. In case of

consistent historic ove rfertilization, a paradigm shift in N management research is needed from

optimizing fertilizer N supply to managing N-saturated soils.

Nitrogen omission in aerobic rice-wheat and wheat-aerobic rice cropping systems in

Mencheng County caused a marked decline in aboveground biomass and grain yield of both

aerobic rice and winter wheat, whereas P and K omission had less effect. The relatively strong

yield reduction in the absence of fertilizer N indicates that, at this site, the native soil N supply is

inadequate to meet crop requirements. Conversely, the low response to P and K omissions of

aerobic rice in the first crop cycle may indicate a relatively adequate native soil P and K supply.

However, P and K became yield limiting for winter wheat in both the first and the second cycle,

because of a higher demand for nutrients by winter wheat than aerobic rice. Calculated for the

whole system, grain yield decreased by 44%, 11% and 7% in the AR-WW sequence, and by 30%,

6% and 0% in the WW- AR sequence due to N, P and K omission.

India. The genotypes Pusa Rice Hybrid 10, Proagro6111 (hybrid), Pusa834 and IR55423-01 (Apo)

produced a yield of more than 4 t ha-1 under aerobic production system irrigated at 40 kPa soil

water tension in the root zone. Among the yield components, biomass, number of grains m-2, and

number of grains ear-1 were most affected by increasing soil water tensions (data not shown).

Compared with typical amounts of water applied to lowland rice fields (under alternate wetting and

drying), the amounts applied in the aerobic rice experiments of 780-1324 mm (irrigation plus

effective rainfall) translated into water savings of the order of 30-40% for production levels of 4-

4.5 t ha-1. Compared with China, yields in the India experiments were lower and the soil wetter as

evidenced by the lower soil water tension (0-40 kPa). The fertilizer application rates in the


Page | 57

experiments were 150 kg N ha-1, 60 kg P ha-1, and 40 kg K ha-1. Further research is needed to find

out if yields can be increased by changing fertilizer management.

Philippines dry season. Except for two experiments at Munoz-PhilRice in 2005with maximum yields

of 5.4-6.1 t ha-1, maximum yields of variety Apo were in the range of 2.9-3.8 t ha-1. These yields

are much lower than the yields in China while the soil water tensions in the root zone were much

lower (usually in the 0-30 kPa range), indicating much wetter soil conditions. There was hardly any

effect of irrigation water application rate because shallow perched water tables (reaching up and

into the root zone) developed underneath the aerobic rice fields. Most of the experiments were

conducted in typical lowland rice environments with relatively shallow groundwater tables and poor

internal drainage of the soils. Therefore, high yields were realized with as little as 274-590 mm

irrigation water input (with groundwater providing a lot of the required water to keep soil water

tensions within the 0-30 kPa range and to meet crop evaporative demand). Yield responded

positively to fertilizer N applications in all experiments, with an application of 120 kg ha-1 sufficient

to reach maximum yields.

Philippines wet season.. Like in the dry season experiments, soil water tensions in the root

zone were generally low, between 0-15 kPa, because of heavy rainfall in the wet season and

shallow groundwater tables. Maximum yields were, with 3.9-4.5 t ha-1, usually higher in the wet

season than in the dry season (see above). Fertilizer N addition had a significant positive effect on

yield. At San Ildefonso, highest yields were obtained at 60 kg N ha-1 and above (statistically the

same), and in Dapdap at 120 kg N ha-1 and above (statistically the same). Overall, there was

considerable lodging in the 120 and 150 kg N ha-1 treatments. Heavy winds and strong rains

frequently occur in the Philippines (central Luzon) near the end of the rainy season, and the risk of

lodging needs to be addressed in fertilizer N recommendations. In the two seasons at Munoz-

PhilRice, a late application of 10 kg N ha-1 at flowering produced significantly higher yields.

Otherwise, yield was insensitive to distribution of N during the growing season.

Laos. The few results for Laos (1 year, 2 sites) show the variability in nutrient response

among sites. The same cultivars produced yields of 3-5.1 t ha-1 at one site (Somsanouk) and only

0.8-3.2 t ha-1 at the other site (Houay Khot). More years of experimentation are needed to derive

firm nutrient management recommendations.

6.2.10.1. Simulation modeling

We used the crop growth simulation model ORYZA2000 (Bouman et al., 2001) to extrapolate

experimental findings to wider environments (soil, weather, hydrology) and compute irrigation

water requirements and yield levels under different irrigation management scenarios. Two studies

were done in China (Table 6.2.24); one extrapolating experimental data from Kaifeng, and one

extrapolating experimental data from Beijing to soils of the Yellow River Basin. The Kaifeng study

was reported by Feng et al. (2007) and Bouman et al. (2007). Here, a summary of the simulation

study using the Beijing data is presented, adapted from Xue et al. (submitted to Irrigation

Science).


Page | 58

Table 6.2.24. Sites of modeling activities per country

Activity/

Country

Phil China India Laos Thai

Simulation

modeling

Beijing -> Yellow River Basin,

Kaifeng

6.2.11. Method

First, the experimental data collected near Beijing in 2002-2004 (see “Water and nutrient

dynamics” paragraph) were used to parameterize and evaluate the ORYZA2000 model. The

evaluated variables included total aboveground biomass, biomass of crop organs, leaf area index,

grain yield, and soil water tension. Next, ORYZA2000 was used to simulate the yield, water inputs

(by irrigation and rainfall), and evapotranspiration (ET) of HD297 under nine irrigation regimes, on

five soil types, and for 34 years of historical weather data. ET flows are real water losses that

deplete water from the system (Molden et al., 2003), whereas total water inputs satisfy the ET

needs plus any deep percolation losses and additions to the soil water storage. The irrigation

regimes were zero irrigation, and 75 mm irrigation water applied when the soil water tension at 20

cm depth reached 10, 20, 30, 50, 70, 100, 200, and 500 kPa. Two soils were from our experiment

fields in 2002 and 2003, which were named Exp02 and Exp03, respectively. The other three soils

were a typical silty loam, loam, and sandy loam, with water retention properties taken from

Wopereis et al (1996). These five soils represent typical soils of the Yellow River Basin (YRB) (Zi,

1999).

6.2.12. Results

Figure 6.2.12 gives the results of the simulations for different soils and irrigation regimes. Yields

were quite comparable on all soils and only gradually declined from an average of around 7500 kg

ha-1 with soil water content kept around field capacity, to 6750 kg ha-1 with irrigation applied at

100 kPa soil water tension. Standard deviations caused by differences in weather were about 900

kg ha-1. With irrigation applied at thresholds above 100 kPa, yields decreased faster and started to

vary among the soils. Yields decreased with decreasing water-holding capacity of the soils, from

Exp03 to silt loam, loam, sandy loam, and Exp02. Under completely rainfed conditions, average

yields still ranged from about 2700 kg ha-1 on the poorest soil to 4000 kg ha--1 on the best soil. As

the year-to-year variability in rainfall was not mitigated by irrigation, the standard deviation went

up to some 1900 kg ha-1.

The long-term average rainfall was 477 mm (Figure 6.2.12b, rainfed bar). Irrigation water

inputs varied most among soil types at relatively low thresholds of irrigation application, and

decreased in reverse order of yield with soil type: from Exp02, to sandy loam, loam, silt loam, and

Exp03. To keep the soil at field capacity (irrigation at 10 kPa) required daily irrigations and an

average amount of irrigation water over all soils of 10485 mm. However, the amount of required

irrigation water decreased fast with increasing threshold of soil water tension. At 20 kPa, irrigation


Page | 59

water inputs were already reduced to 823 mm on Exp03 and to 1260 mm on Exp02. At 100 kPa,

irrigation water inputs were as low as 200 mm on Exp03 and 320 mm on Exp03.

Differences in evapotranspiration (ET) among the soils were minimal with low irrigation

thresholds, and slowly increased with increasing thresholds (Figure 6.2.12c). With a continuously

wet top soil with irrigation applied at 10 kPa, total ET was relatively high at around 690 mm. ET

dropped fast to around 510 mm with irrigation applied at 20 kPa to 370-416 mm under rainfed

conditions.

Both water productivities WPIR and WPET showed a maximum value with intermediate

irrigation regimes. The WPIR increased from as low as 0.07 g kg-1 at 10 kPa irrigation threshold to

0.89-1.05 g kg-1 at 200 kPa because the decrease in irrigation water requirements over that

range outweighed the decrease in yield (Figure 6.2.12d). Beyond 200 kPa, however, the yield

decrease outweighed the reduced irrigation water requirements, and WPIR declined again to 0.55-

0.82 g kg-1 under rainfed conditions. The differences in WPIR among soil types were relatively

pronounced, and showed the same trends as those in yield. The WPET was highest and stable at

about 1.45 g kg-1 between irrigation thresholds of 20 to 100 kPa, and then declined gradually to

0.70-0.93 g kg-1 under rainfed conditions (Figure 6.2.12e).

Figure 6.2.12. Average simulation results over 34 years as a function of threshold level of soil

water tension for irrigation on five soils: yield (a), rainfall plus irrigation (b), evapotranspiration

(c), water productivity based on total water input WPIR (d), water productivity based on

evapotranspiration WPET (e). The error bars indicate the standard deviation caused by weather.

Yield (kg ha-1)

0

1500

3000

4500

6000

7500

9000

10 20 30 50 70 100 200 500 RainfedThreshold of soil w ater tension (kPa)

EXP03 SILOAM LOAMSLOAM EXP02

A


Page | 60

Rainfall and irrigation (mm)

0

400

800

1200

1600

2000

20 30 50 70 100 200 500 RainfedThreshold of soil w ater tension (kPa)


B

Evapotranspiration (mm)

0

100

200

300

400

500

600

700

800



C

WPIR (g kg-1)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8



D


Page | 61

WPET (g kg-1)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

10 20 30 50 70 100 200 500 Rainfed

Threshold of soil w ater tension (kPa)


E

Average maximum simulated yields were 6.8-7.5 t ha-1 with soil water potentials in the root zone

staying between 0 and 100 kPa. These yields were 0.8-1.5 t ha-1 higher than the maximum of 6 t

ha-1 in our field experiments. There may be at least 3 reasons for this difference. First, in nearly all

of the experimental treatments, there were periods when the tensiometers failed to record

because the threshold of 100 kPa was exceeded, so that higher soil water tensions occurred that

could have reduced yields. When simulated soil water tensions were allowed to reach 200 kPa,

yields already dropped to 6-6.1 t ha-1 (Figure 6.2.12a). Second, aerobic rice is a relatively new

crop and best management practices still need to be developed. Therefore, the management

applied in the field experiments may not have resulted in the highest possible yields. Third, the

variety HD297 might have been sink limited under high-yielding conditions. Analyzing yield

formation in HD297, Bouman et al. (2006) reported that the source strength (quantified by light

use efficiency) of this variety was comparable to that of lowland varieties with yield potentials of

more than 8 t ha-1. The panicle size of HD297, however, may be too small to attain such yields

(visual observation by authors).

Simulated yields sharply decreased with irrigation water application above the 100 kPa soil

water tension threshold. But even under completely rainfed conditions, yields were still 2.7-4 t ha-

1, depending on soil type. Required water inputs declined sharply with irrigation thresholds

increasing from 10 to 50 kPa, and then declined slowly with increasing thresholds from 50 kPa

onwards. Since yield followed the opposite trend with irrigation threshold, the water productivity

WPIR showed a maximum at intermediate irrigation thresholds (Figure 6.2.12d). When water

resources are limited, the best irrigation scheme would optimize water productivity rather than

grain yield. In our simulations, the highest WPIR was obtained at irrigation thresholds of 100-200

kPa, with yields of 6-6.8 t ha-1 and only 112-320 mm irrigation water applied (and 477 mm

rainfall). In farmers’ practice such irrigation amounts would translate into 2-4 irrigation

applications of 50-75 mm each, which is about the irrigation frequency of early adopters of aerobic

rice in the Kaifeng area (Bouman et al., 2007). Compared with other upland crops in the YRB, this

irrigation water requirement is lower than that for winter wheat (Sun et al., 2006) but higher than

that for summer maize (Wang et al., 2001).

Evapotranspiration (ET) was quite stable over a wide range of irrigation thresholds, and

only declined slightly from about 690 mm at 20 kPa to 370-416 mm under purely rainfed

conditions. At 20 kPa, the total water inputs were 1300-1739 mm (depending on soil type), and


Page | 62

hence some 610-1049 mm of water left the root zone by deep percolation. The amount of deep

percolation gradually decreased to

61-107 mm under purely rainfed conditions. Although deep percolation is a loss to farmers, it re-

enters the hydrological cycle and is potentially available for reuse, for example by groundwater

pumping (Hafeez et al., 2007). Therefore, for irrigation system managers, the optimum irrigation

scenario could be when the highest water productivity with respect to evapotranspiration is

obtained (Loeve et al., 2004). In our simulations, highest levels of WPET were realized with

irrigation thresholds between 10 and 100 kPa (Figure 6.2.12e). Irrigation thresholds of 100-200

kPa therefore seem a suitable balance between the interests of farmers (striving for high WPIR)

and irrigation system managers (striving for high WPET).

Soil type had relatively little effect on yield except under purely rainfed conditions, with the

lighter-textured soil with lowest soil water holding capacity having lowest yields. Soil type had the

strongest effect on irrigation water inputs with irrigation application thresholds below 70 kPa. The

effect of soil type was most pronounced on water productivity with respect to total water inputs:

WPIR decreased on soils with decreasing water holding capacity.

6.2.13. Conclusion

Field experiments and simulations show that, on typical freely-draining soils of the YRB, aerobic

rice yields with HD297 can reach up to 6 t ha-1, with, on average some 477 mm rainfall and 112-

320 mm of irrigation water dosed in 2-4 applications to keep the soil water tension in the root

zone below 100-200 kPa. Drought around flowering should be avoided by targeted irrigation

applications to avoid the risk of spikelet sterility that would result in low grain yields. To further

increase yield and productivity of aerobic rice in the YRB, we suggest the following research

activities:

1. “Management”. Optimizing the productivity of aerobic rice in a cropping system context.

Studies should address the potential to increase yields through improved crop

management practices such as establishment (such as optimum seed density, row

spacing) and nutrient management (macro and micro nutrients). Moreover, such studies

should look at the potential role of aerobic rice in multiple cropping systems and crop

rotations (such as optimizing sowing dates, crop durations, cropping calendar), and should

address whole cropping system productivity.

2. “Germplasm improvement”. The development of new varieties should focus on the

potentials for yield increase through increasing the sink size and harvest index of the

current variety HD297.

3. “Model improvement”. The current version of ORYZA2000 captures the effect of water

stress at flowering through the effect of increased canopy temperature on spikelet sterility

only (Bouman et al., 2001). More insight in the mechanisms of spikelet sterility should be

obtained and built in ORYZA2000, especially at soil water tensions above 100 kPa. Also,

ORYZA2000’s water balance model should be specifically tested in dry soil conditions

where the soil water potential exceeds 100 kPa. The presented model scenarios used


Page | 63

weather data from Beijing and freely-draining soils only. Next steps would be to study the

effect of groundwater tables on yield and water requirements, and to include weather data

from other sites on the YRB.

6.3. Objective 3: Sustainability

Yield decline under continuous monocropping of upland rice has been reported in Japan, Brazil,

and the Philippines (Nishizawa et al. 1971; Pinheiro et al, 2006; Ventura and Watanabe 1978;

George et al. 2002). Yield decline of continuous upland rice is generally believed to be caused by

soil sickness, which may include the buildup of nematodes (Nishizawa et al. 1971) or soil

pathogens (Ventura et al. 1981), changes in nutrient availability in the soil (Lin et al. 2002), or

growth inhibition by toxic substances from root residues (Nishio and Kusano 1975). So far, the

causes of yield decline in continuous upland rice are still unknown, and we don’t know if the same

yield decline will occur in continuous aerobic rice systems. Documenting any yield decline in

continuous aerobic rice, and understanding its causes, is necessary to develop sustainable

management strategies of aerobic rice systems. Here, we present our work done in the Philippines

with pot and field experiments (Table 6.3.1).

Table 6.3.1. Sites of sustainability research activities per country

Activity/country Philippines China India Laos Thailand

Field experiment Los Banos,

Dapdap

Pot experiment Los Banos

6.3.1. Gradual yield decline

In 2001, IRRI established a long-term field experiment to compare the agronomic performance of

aerobic and flooded rice. The field experiment was conducted at the International Rice Research

Institute (IRRI) farm at Los Baños, Laguna, Philippines (14°11′N, 121°15′E, 21 m asl) in both dry

season (DS, January-May) and wet season (WS, June-October). There were three main fields,

subdivided into plots with different treatments, and four replications. Up to 2004, there were three

water treatments in the main fields: aerobic rice in both DS and WS, flooded rice in both DS and

WS, and aerobic in DS and flooded rice in WS. The continuous aerobic and flooded treatments

were divided into a “full N” treatment and a “0-N treatment”. Full N was 150 kg urea-N ha-1 in the

DS and 70 kg urea-N ha-1 in the WS. In all fields, phosphorus (60 kg P ha-1 as solophos),

potassium (40 kg K ha-1 as potassium chloride), and zinc (5 kg Zn ha-1 as zinc sulfate

heptahydrate) were incorporated one day before transplanting. From 2004 onward, changes were

introduced to study effects of crop rotations, fallowing, and conversion of flooded fields into “fresh”

aerobic fields. The continuous aerobic rice cropping continued till 2008. Variety “Apo” (PSBRc9)

was used throughout the experiment. Flooded plots were puddled and kept continuously flooded


Page | 64

with 5-10 cm of water depth from transplanting until 2 weeks before harvest. The aerobic plots

were dry-ploughed and harrowed but not puddled during land preparation. One day before

transplanting, aerobic plots were soaked with irrigation water overnight to facilitate transplanting.

Transplanting was done for aerobic rice to keep seedling density constant across seasons.

Afterward, aerobic plots were flash irrigated with about 5 cm water each time only when the soil

moisture tension at 15 cm depth reached –30 kPa. Around flowering, the threshold for irrigation

was reduced to –10 kPa to prevent spikelet sterility.

In our project, we summarized the results over the period 2001-2004 (1), initiated follow-

up studies on understanding the causes of yield decline with a focus on nutrient supply (2), studied

“restoration” attempts in the long-term field experiment by crop rotations (3), and continued

monitoring yields under continuous aerobic conditions in the period 2005-2007 (4). In the next

sections, we present results for each of these four activities.

Peng et al. (2006) and Bouman et al. (2005) presented the key findings of the long-term field

experiment with respect to yield and water use. Here, we summarize the findings on yield decline

and present some additional data on nematode counts. Aerobic rice yields were higher in the DS

than in the WS (Figure 6.3.1). In the DS, yields were lower in 2003-2004 than in 2001-2002,

though the difference was only small. In the WS, yields were 3.5-4 t ha-1 without any clear

pattern over the years. When compared with the flooded yields, the yields under aerobic

conditions declined consistently from 2001 to 2004, both the DS and in the WS. However, it is

unclear whether this relative decline is a consequence of increasing flooded yields over time or

decreasing aerobic yields over time.

In 2004, the flooded plots in the previous six seasons were converted to “fresh” aerobic plots,

while flooded plots only in WS became flooded in 2004 DS. This change allowed a direct

comparison between rice grown under aerobic conditions in the soil where flooded rice has been

grown continuously in previous seasons (first season aerobic rice) and in the soil where aerobic

rice has been grown continuously in previous six seasons (seventh season aerobic rice). Here, the

effect of continuous cropping is very clear: the yield in the same year in the “fresh” aerobic field is

some 2.5 t ha-1 higher than in the seventh-season continuous aerobic field, while the yield in the

“fresh” aerobic field is only 1.5 t ha-1 less than in the flooded field (Table 6.3.2).


Page | 65

Table 6.3.2. Grain yield, total biomass, and harvest index of Apo grown under aerobic conditions in

the soil where flooded rice has been grown continuously in previous seasons (1st season aerobic

rice) or in the soil where aerobic rice has been grown continuously in previous six seasons (7th

season aerobic rice) in comparison with flooded rice in the dry season of 2004 at IRRI farm.

Parameters 1st season

aerobic rice

(A1st)

7th season

aerobic rice

(A7th)

Flooded

rice

(F)

Grain yield (t ha-1) 6.32 b 3.77 c 7.78 a

Total biomass (g m-2) 1343 b 862 c 1604 a

Harvest index (%) 45.3 b 45.4 b 47.1 a

Within a row, means followed by different letter are significantly different at 0.05 probability level

according to Least Significant Difference (LSD) test.


Page | 66

Figure 6.3.1. Grain yield of Apo grown under aerobic and flooded conditions and the yield

difference between aerobic and flooded rice in dry seasons (A) and wet seasons (B) of 2001-04 at

IRRI farm. The aerobic yields include all fields that were aerobic in both the aerobic-aerobic and

the aerobic-flooded treatments. Source: Peng et al., 2006.

Fig. 1

2001 2002 2003 2004

Yie

ld (

t ha-1

)

1.0

2.5

4.0

5.5

7.0

8.5

Diff

eren

ce (

%)

10

20

30

40

50

60

70

80

Aerobic Flooded Difference

A

Year

2001 2002 2003 20041.0

2.5

4.0

5.5

7.0

8.5

10

20

30

40

50

60

70

80B

We measured nematodes (Meloidogyne graminicola) in the roots of the stubbles within two weeks

after harvest (Table 6.3.3). Before our experiment started, nematode numbers in the stubble of

flooded rice (in all plots) of the previous experiment were 2-9 g-1 fresh root weight. In the

continuous aerobic fields, the number of nematodes increased rapidly with a peak in the 2002 and

2003 DS. The aerobic-flooded treatment was not able to suppress nematodes in the aerobic phase

whereas in the flooded phase, nematode counts were very low again. In the continuous flooded

fields, nematode counts remained low except for the 2002 and 2003 DS where moderate levels

were counted. Although nematode counts were clearly high in the aerobic fields, no statistical

correlation was found between nematode count and yield level over time. Moreover, whereas the

“fresh” aerobic field in the 2004 DS had a high yield of 6.3 t ha-1 and the seventh-season a low

yield of 3.8 t ha-1, the number of nematodes was higher in the fresh aerobic field (658) than in the

seventh-season aerobic field (368), though because of high spatial variability these differences

were not significant.


Page | 67

Table 6.3.3. Number of Meloidogyne graminicola nematodes per gram fresh root weight after

harvest, in Apo grown under aerobic (A) and flooded (F) conditions, IRRI farm 2001-2004.

Continuous

aerobic

Alternate

aerobic-flooded

Continuous flooded

Aerobic in 2004 DS

Baseline

flooded

6 9 2

Aerobic DS

Aerobic WS

Aerobic DS

Flooded WS

Flooded DS

Flooded WS

2001 DS 875 (A) 279 (A) 6 (F)

2001WS 491 (A) 11 (F) 4 (F)

2002 DS 2530 (A) 1760 (A) 134 (F)

2002 WS 1499 (A) 27 (F) 34 (F)

2003 DS 2089 (A) 3054 (A) 380 (F)

2003 WS 694 (A) 51 (F) 45 (F)

2004 DS 368 (A) 35 (F) 658 (A)

6.3.2. Yield restoration

From 2004 onward, restoration attempts were made by introducing periods of flooding, fallowing,

and different crops into the long-term aerobic rice field experiment (Table 6.3.4).

1. Flooding: In 2004 DS and WS, the alternate aerobic-flooded fields (aerobic rice in DS and

flooded rice in WS) were converted into flooded rice. In the 2005 DS, aerobic rice was

planted again to create a 1st season aerobic rice after three seasons of flooded rice. Full N

rates were applied.

2. Fallowing: In 2004 DS and WS, half of the continuous aerobic rice fields were converted to

fallow. In the 2005 DS, aerobic rice was planted again to create a 1st season aerobic rice

after two seasons of fallow. Full N rates were applied. The remaining half was kept

continuously cropped to aerobic rice since 2001 DS.

3. Crop rotation: In 2004 the flooded fields from 2001-2003 were cropped to aerobic rice in

both DS and WS. In 2005, they were reverted back to flooded rice to serve as reference

for aerobic rice in other fields. In 2006, different upland crops were planted in the Ds and

WS: soybean, maize, sweet potato. In addition, a fallow was included. After harvest, all

above-ground material was removed from the plots. In 2007DS, aerobic rice was planted

again. Full N rates were applied.


Page | 68

Table 6.3.4. Cropping pattern in long-term aerobic rice experiment at the IRRI farm, 2002-2007.

Main field I Main field II Main field III

2001 DS Aerobic Aerobic Flooded

2001WS Aerobic Flooded Flooded


2002 WS Aerobic Flooded Flooded


2003 WS Aerobic Flooded Flooded

2004 DS Fallow Aerobic Flooded Aerobic

2004 WS Fallow Aerobic Flooded Aerobic

2005 DS Aerobic Aerobic Aerobic Flooded

2005 WS Aerobic Aerobic Aerobic Flooded

2006 DS Aerobic Soy

bean

Maize Sweet

potato

Fallow

2006 WS Aerobic Soy

bean

Maize Sweet

potato

Fallow

2007 DS Aerobic Aerobic Aerobic Aerobic Aerobic

2007 WS Aerobic Aerobic Aerobic Aerobic Aerobic

The effect of fallowing and flooding on yield of aerobic rice is given in Table 6.3.5. In the N-

fertilized fields, the effect of both two seasons fallowing and three season of flooding was

significantly positive: yield increased by about 1.5 t ha-1 in the fist season and 1 t ha-1 in the

second (relative to the long-term aerobic rice fields). There was no difference in yield between the

fallowing and flooding restorations. In the 0-N fertilized fields, there was no significant effect of

both the fallowing and the flooding restoration: yields were statistically the same as in the long-

term aerobic rice fields.


Page | 69

Table 6.3.5. Effect of two seasons of fallow and three seasons of flooding on the yield of Apo in

aerobic and flooded fields in 2005DS and 2005WS.

Treatment Yield (t ha-1)

With N Without N

2005 DS

9th Aerobic 3.03 c 2.68 b

1st Aerobic after fallowing 4.43 b 2.50 b

1st Aerobic after flooding 4.51 b 2.89 b

Flooded 6.79 a 4.86 a

2005 WS

10th Aerobic 3.18 c 2.57 c

2nd Aerobic after fallowing 4.27 b 3.02 b

2nd Aerobic after flooding 4.01 b 2.82 bc

Flooded 5.71 a 5.03 a

Within a column, means followed by different letter are significantly different at 0.05 probability

level according to Least Significant Difference (LSD) test.

The effect of fallowing and different crop rotations on yield of aerobic rice is given in Table 6.3.6.

In the first season, aerobic rice yields after the three upland crops were statistically the same, and

about 0.8-1.1 t ha-1 higher than after fallowing. Compared with all the previous years, aerobic rice

yields after upland crops were relatively high at 5.4-5.8 t ha-1. Surprisingly, even the yield in the

continuous aerobic rice fields was relatively high and comparable with the yields after the three

upland crops (no statistical analysis done yet).


Page | 70

Table 6.3.6. Effect of two seasons of fallow and different upland crops on the yield of Apo in

aerobic fields in 2007DS.

Treatment Yield (kg ha-1)

2007 DS

13th Aerobic (0-N) 3.02

13th Aerobic (full N) 5.19

1st Aerobic after fallowing 4.63 b

1st Aerobic after maize 5.41 a

1st Aerobic after sweet potato 5.41 a

1st Aerobic after Soy bean 5.76 a

6.3.3. Long-term aerobic rice yields

Tables 6.3.7a,b give the yields in the continuous aerobic rice fields and in the flooded rice fields

from 2001 to 2007. In the DS, the yield under aerobic conditions consistently declined compared

with the yield under flooded conditions from 2001 to 2005. Also in absolute terms, the continuous

aerobic rice yields declined, under both N-fertilized and 0-N conditions. However, in both 2006 and

2007, aerobic rice yields spectacularly increased and even attained the highest recorded yield of

5.2 t ha-1 under N-fertilized conditions in 2007. Unfortunately, there were no flooded treatments to

compare with. In the WS, no consistent yield decline was found over the period 2001-2007.

In both the DS and the WS, the sustainability of aerobic rice under 0-N conditions is

remarkable: aerobic rice yield in the DS was still 3 t ha-1 after 13 crops, and 2.4 t ha-1 in the WS

after 12 crops.

Table 6.3.7a. Yield of Apo under continuous aerobic and under flooded conditions, and ratio of

aerobic over flooded yield, IRRI farm, 2001-2007 DS.

2001DS 2002DS 2003DS 2004DS 2005DS 2006DS 2007DS

Aerobic +N 4.37 5.06 3.84 3.78 3.05 4.45 5.19

Aerobic –N 3.08 3.29 3.16 2.73 2.47 2.47 3.02

Flooded +N 5.06 7.33 6.81 6.33 6.79 - -

Flooded –N 3.63 3.81 4.06 4.02 4.86 - -

A/F +N (%) 86 69 56 60 45 - -

A/F -N (%) 85 86 78 68 51 - -


Page | 71

Table 6.3.7b. Yield of Apo under continuous aerobic and under flooded conditions, and ratio of

aerobic over flooded yield (A/F), IRRI farm, 2001-2007 WS.

2001WS 2002WS 2003WS 2004WS 2005WS 2006WS 2007WS

Aerobic +N 3.6 3.49 4.2 3.36 3.18 3.16 -

Aerobic –N 2.59 3.13 3.74 2.54 2.57 2.36 -

Flooded +N 4.57 4.77 6.04 4.22 5.71 - -

Flooded –N 3.65 3.24 4.47 3.06 5.03 - -

A/F +N (%) 79 73 70 80 56 - -

A/F -N (%) 71 97 84 83 51 - -

6.3.4. Causes for yield decline

A large number of pot and micro-plot experiments were done with the soil from the long-term

aerobic rice experiment at IRRI to determine the causes of the yield decline and propose remedial

measures. Pot experiments were done in greenhouses and screenhouses at IRRI, and variety Apo

was used in all experiments.

Oven heating

Soil sterilization by oven heating to remedy “soil sickness” (Anderson and Magdoff 2005;

Kirkegaard et al. 1995; Sasaki et al. 2006). Six pot experiments with different soil heating

treatments were conducted at the IRRI farm using soil collected from the top 25 cm of three fields

that were previously grown to aerobic rice or to flooded rice. Details are reported by Lixiao Nie et

al. (2007), and here we summarize the main findings. Soils were taken from fields where aerobic

rice was grown continuously for 10 and 5 seasons, and from fields where flooded rice was grown.

Soil heating was done at 120 °C for 12 hours.

Oven heating of soil with an aerobic history increased rice plant growth significantly over

the unheated control (Table 6.3.8). The response of plant growth to soil heating was highest in

leaf area, followed by total biomass and stem number. Although soil heating does not reveal the

cause of soil sickness and cannot differentiate between biotic (e.g., soil-borne pathogens) or

abiotic (e.g., soil chemical) causes of soil sickness, the evidence from the various pot experiments

(not shown here, see Lixiao Nie et al., 2007) suggests that abiotic factors are more likely to cause

“soil sickness” in our field experiment.


Page | 72

Table 6.3.8. Plant growth of Apo grown aerobically without soil heating (control) and after soil

heating using aerobic and flooded soils from IRRIs’ long-term aerobic rice experiment. Source:

Lixiao Nie et al., 2007.

Parameter Control Heated Increase (%)

10th-season aerobic soil

Stem number per pot 3.2 b 19.8 a 519

Plant height (cm) 37.4 b 67.2 a 80

Leaf area (cm2 pot-1) 60 b 991 a 1552

Total biomass (g pot-1) 0.6 b 7.5 a 1150

5th-season aerobic soil

Stem number per pot 3.0 b 10.4 a 247

Plant height (cm) 35.8 b 57.8 a 61

Leaf area (cm2 pot-1) 66 b 394 a 497

Total biomass (g pot-1) 0.7 b 3.1 a 343

Flooded soil

Stem number per pot 9.0 a 10.4 a 16

Plant height (cm) 57.0 a 56.0 a -2

Leaf area (cm2 pot-1) 296 a 356 a 20

Total biomass (g pot-1) 2.6 a 3.5 a 35

Within a row, means followed by different letters are significantly different at 0.05 probability level

according to least significant difference (LSD) test.

Nutrient supply

Soil oven heating was shown above to alleviate soil sickness caused by continuous copping of

aerobic rice. Heating can kill pathogenic nematodes, fungi, and bacteria, but can also facilitate the

release of nutrients from the soil by enhancing mineralization or transforming nutrients into more

available forms. Using micro-plots and pot experiments, we explored the effects of N, P, K and

micronutrients on growth and yield of aerobic rice grown in soil from IRRI’s long term aerobic rice

experiment. Details of the experiments and results are reported by Lixiao Nie et al. (in prep 1),

and here we summarize the key findings.

Micro-plots of 1.0 × 1.0 m were established in 9th-season aerobic rice field (2005 DS) of

IRRI’s long-term aerobic rice experiment. The treatments were Yoshida nutrient solution (Yoshida

et al., 1976), Yoshida solution without NPK, and control. In the micro-plots with full Yoshida

solution, the total amount of elements received was 144 kg N, 36 kg P, 144 kg K, 188 kg Ca, 188


Page | 73

kg Mg, 2.35 kg Mn, 0.235 kg Mo, 0.94 kg B, 0.047 kg Zn, 0.047 kg Cu, and 9.4 kg Fe per hectare.

The application of Yoshida solution without NPK increased grain yield over the control, but the

difference was not significant (Table 6.3.9). The application of the full Yoshida solution, however,

increased yield by 47% and 24% over the control and Yoshida solution without NPK, respectively.

These results suggest that the combination of N, P, and K rather than micronutrients alleviated soil

sickness caused by continuous cropping of aerobic rice.

Table 6.3.9. Biomass, yield, and yield components of Apo grown with Yoshida nutrient solution

without NPK and with complete Yoshida solution, in comparison with control, IRRI farm, 2005 DS.

Parameter Control Yoshida solution

without NPK

Yoshida

solution

Grain yield (t ha-1) 3.18 b 3.77 b 4.69 a

Aboveground biomass (g m-2) 809 b 903 b 1067 a

Harvest index (%) 34.8 a 37.2 a 39.2 a

Panicles m-2 254 b 262 b 304 a

Spikelets panicle-1 87.0 b 94.9 ab 99.3 a

Spikelets m-2 (x103) 22.0 c 24.8 b 30.2 a

Grain filling (%) 68.5 a 71.4 a 73.6 a

1000-grain weight (g) 18.6 a 18.9 a 18.8 a



In pot experiment 1, soil was collected from the 11th-season aerobic rice field and from a flooded

field. For both the aerobic and the flooded soil, the treatments were an unfertilized control, oven

heating at 120ºC for 12 hours, and nutrient additions of 0.90 g N as urea, 0.64 g P as solophos,

and 1 g K as potassium chloride per pot. Pot experiment 2 used the same soils, and had a control,

1 oven-heating, and five nutrient treatments: 0.90 g N as urea (1); 0.32 g P as solophos (2); 2 g

K as potassium chloride (3); Yoshida solution without NPK (4); 0.90 g N as urea, plus 0.32 g P as

solophos, plus 2 g K as potassium chloride, plus Yoshida solution without NPK (all values per pot).

The total amount of nutrients in Yoshida solution without NPK was 96 mg Ca, 96 mg Mg, 1.2 mg

Mn, 0.12 mg Mo, 0.48 mg B, 0.024 mg Zn, 0.024 mg Cu, and 4.8 mg Fe per pot. In both pot

experiments, Apo was grown aerobically in all pots.

In pot experiment 1, the application of N, P, and K improved plant growth and leaf N

nutrition significantly over the control in both aerobic and flooded soils (Table 6.3.10). However,

the plant response was greater in the soil from the aerobic field than in the soil from the flooded

field. This confirms that aerobic soil used in this study was “sick” compared with flooded soil, and


Page | 74

that the addition of NPK was effective to reduce the soil sickness. Oven heating also improved

plant growth over the control, in both soils. Again, the response was greater in aerobic soil than in

flooded soil. In pot experiment 2, both urea and solophos increased plant growth significantly over

the control (Table 6.3.11). The application of Yoshida solution and NPK, and soil heating, also

increased plant growth, and the effect was larger than in the urea and solophos treatments. The

application of Yoshida solution without NPK and the application of potassium did not consistently

increase plant growth over the control. The results eliminated the possibility of K in alleviating soil

sickness.

Table 6.3.10. Plant growth of Apo grown with NPK application and under oven heating in

comparison with the control, pot experiment 1, IRRI.

Parameter Control NPK Oven heating

Soil from aerobic field

Plant height (cm) 41.7 b 63.5 a 63.8 a

Stem number per pot 7.5 c 25.0 b 38.8 a

Leaf area (cm2 pot-1) 72 c 729 b 1306 a

Total biomass (g pot-1) 1.74 c 11.13 b 23.70 a

SPAD value 30.5 b 40.1 a 40.9 a

Soil from flooded field

Plant height (cm) 57.3 b 67.8 a 65.7 a

Stem number per pot 20.5 b 26.8 a 28.0 a

Leaf area (cm2 pot-1) 479 b 923 a 854 a

Total biomass (g pot-1) 9.24 b 14.31 a 14.38 a

SPAD value 37.4 b 41.6 a 40.6 a




Page | 75

Table 6.3.11. Plant growth of Apo under different nutrient treatments and oven-heating, in

comparison with the control, pot experiment 2, IRRI.

Treatment Plant height

(cm)

Stem number

pot-1

Leaf area

(cm2 pot-1)

Total biomass (g

pot-1)

Control 57.8 e 6.8 d 318 d 3.8 d

Urea 72.2 b 9.2 c 675 c 5.3 c

Solophos 66.8 c 9.3 c 604 c 6.0 c

Potassium chloride 61.0 d 6.3 d 302 d 3.2 d

Yoshida minus NPK 57.5 e 6.7 d 342 d 4.1 d

(1) + (2) + (3) + (4) 75.8 a 11.7 b 933 b 7.8 b

Oven heating 76.5 a 13.5 a 1059 a 10.9 a

Within a column, means followed by different letters are significantly different at 0.05 probability

level according to Least Significant Difference (LSD) test.

We concluded that micronutrients were not effective in increasing plant growth in pot experiments

nor of increasing grain yield in field micro-plot experiments. Plant growth was not improved with

the application of K or of P fertilizers (Ca-Mg phosphate and rock phosphate) and P chemical

reagents (monosodium phosphate dihydrate) (Pot experiments and data not shown; Lixiao Nie et

al. in prep 1). However, slight growth increase was observed with the application of calcium

superphosphate, and large growth increase with solophos. Solophos and calcium superphosphate

contained 2.9% and 1.7% N, respectively, and we can not rule out the possibility that the added N

in these P fertilizers contributed to improved crop performance. From other pot experiments (data

not shown; Lixiao Nie et al. in prep 1), however, it was concluded that P nutrition was not

associated with the soil sickness in IRRI’s continuous aerobic rice experiment.

Nitrogen form

From the nutrient experiments reported above, it was concluded that N application may alleviate

the effects of soil sickness in IRRI’s continuous aerobic rice experiment. Follow-up pot experiments

were conducted using soil collected from the 11th-season aerobic rice field and from a flooded field

of IRRI’s long-term aerobic rice experiment (Lixiao Nie et al. in prep 2). In all experiments, Apo

was grown aerobically in all pots. In experiment 1, we used only soil from the aerobic fields.

Beside an oven-heated treatment, we had 25 nutrient treatments consisting of 5 forms of N

chemical reagents (ammonium sulfate ((NH4)2SO4), urea (CO(NH2)2), ammonium nitrate

(NH4NO3), ammonium chloride (NH4Cl), and potassium nitrate (KNO3)) by 5 N rates (0, 0.3, 0.6,

0.9, and 1.2 g N/pot). In experiment 2, we used both soils. Beside an oven-heated treatment, we

had 2 nutrient treatments: 1.2 g N/pot as urea, and 1.2 g N/pot as ammonium sulfate.


Page | 76

In experiment 1, increasing N rates increased plant growth in the aerobic soil, when N was applied

as ammonium sulfate and urea (Table 6.3.12). Compared with urea, ammonium sulfate was more

effective in promoting plant growth. Plant growth was not consistently improved by the application

of ammonium nitrate. Though most growth parameters increased when N was applied as

ammonium chloride, total biomass did not respond significantly. Potassium nitrate had a negative

effect on plant growth, even to the extent that plants died at the rate of N4. In experiment 2, both

ammonium sulfate and urea application significantly improved plant growth in the aerobic soil

(Table 6.3.13). However, the effect of ammonium sulfate was larger than that of urea. Urea had

no effect on plant growth in the soil from flooded fields, whereas the effect of ammonium sulfate in

the soil from flooded fields was much smaller than that in the soil from aerobic fields. Oven-

heating of the soil increased plant growth in both soils.

Table 6.3.12. Aboveground biomass (g pot-1) of Apo grown under five N forms with five N rates

and soil oven-heating treatment in pot experiment 1, IRRI.

N rates (NH4)2SO4 NH4Cl CO(NH2) 2 NH4NO3 KNO3

N0 3.42 d 3.42 bc 3.42 b 3.42 b 3.42 b

N1 4.44 d 2.88 c 3.89 b 3.13 b 1.90 c

N2 7.03 c 3.80 bc 4.55 b 2.98 b 1.77 c

N3 7.19 c 5.14 b 4.57 b 2.55 b 0.60 d

N4 11.90 b 4.81 b 4.82 b 2.84 b 0.00 d

Oven heating 15.38 a 15.38 a 15.38 a 15.38 a 15.38 a

Within a column under each parameter, means followed by different letters are significantly

different at 0.05 probability level according to Least Significant Difference (LSD) test.


Page | 77

Table 6.3.13. Plant growth of Apo grown under urea and ammonium sulfate application and soil

oven-heating in comparison with the untreated control in pot experiment 2, IRRI.

Parameter Control Urea Ammonium sulfate Oven-heated

Plant height (cm) 45.2 c 58.7 b 77.5 a 80.8 a

Stem number per pot 4.7 d 8.8 c 19.7 b 22.2 a

Leaf area (cm2 pot-1) 105 d 300 c 951 b 1423 a

Root dry weight (g pot-1) 0.27 d 0.72 c 1.52 b 2.73 a

Total biomass (g pot-1) 1.23 d 3.41 c 10.03 b 16.71 a

Plant height (cm) 68.8 c 74.3 b 75.7 ab 79.2 a

Stem number per pot 11.5 b 14.3 b 21.3 a 20.7 a

Leaf area (cm2 pot-1) 589 b 690 b 923 a 1106 a

Root dry weight (g pot-1) 1.48 ab 1.00 c 1.23 bc 1.66 a

Total biomass (g pot-1) 6.92 c 7.27 bc 9.32 ab 10.87 a



We concluded that the application of nitrogen could reverse the decline in crop growth in

continuously cropped aerobic rice in IRRI’s long-term field experiment. Though in the long-term

experiment, 150 and 70 kg urea-N ha-1 are continuously applied in the DS and WS, respectively,

the extra N applications significantly improved the plant growth in our pot experiments.

Ammonium sulfate was much more effective on improving plant growth than urea or other sources

of N.

6.3.5. Yield “collapse”

Beside the gradual yield decline under continuous cropping of aerobic rice reported above, we

encountered two cases of “immediate yield collapse” in fields cropped to aerobic rice for the very

first time. One field experiment at Dapdap, Tarlac, (same site as where other successful

experiments on aerobic rice were conducted, see paragraphs on “Water and nutrient dynamics”

and on “Crop establishment”) and one at the IRRI farm, on water x nitrogen interaction showed

complete yield failures. Though our study on “yield collapse” is not part of the CPWF-project, we

present the results of the Dapdap experiment here to “flag the problem”.

A water by N experiment was conducted in the dry season 2004 and 2005 at Dapdap

(120.73ºN, 15.62ºE, 26 m asl) in Central Luzon, The Philippines. The soil was loamy sand with

71% sand, 22% silt, and 7% clay in the top soil (15 cm). Before our experiment, the soil was

cropped to rainfed lowland rice under conditions of intermittent flooding (because of irregular


Page | 78

rainfall and high soil water permeability). The water treatments were: irrigation twice per week

(W1), once per week (W2) and once in 2 weeks (W3, modified in 2004 to weekly from panicle

initiation on). During the critical flowering stage all water treatments were irrigated twice per week

for 4 to 5 weeks. Irrigation was applied by sprinklers early in the season and by flash flooding later

in the season. The N treatments were: 0 kg N ha-1 (N1), 60 kg N ha-1 (N2), 120 kg N ha-1 (N3),

160 kg N ha-1 (N4), and 200 kg N ha-1 (N5). In 2005, N5 was cancelled and N4 adjusted to 165 kg

N ha-1. There was complete yield failure on all treatments in both years (Table 6.3.14), even

though up to 200 kg urea-N ha-1 and some 1000 mm water were applied. The investigation of the

possible causes of yield collapse at Dapdap, and in the similar experiment at IRRI (data not

shown), is in full progress and focuses on both biotic (nematodes, root fungi) and abiotic (micro-

and macro-nutrients, soil pH) factors. From the initial analysis of the Dapdap experiment, it is

concluded that the yield collapse was not a direct result of reduced water and/or N availability.

There was good evidence for root knot nematodes and potentially micronutrient imbalances as

causal factors.

Table 6.3.14. Yield (t ha-1), total straw biomass (t ha-1), and harvest index, per water

(W) and nitrogen (N) treatment, Dapdap, 2004 and 2005.

Yield Straw Harvest index

W1 W2 W3 Av W1 W2 W3 Av W1 W2 W3 Av

2004

N1 0.11ab 0.02a 0.03a 0.05a 2.8b 3.0b 3.0a 2.9b 0.07a 0.01a 0.03a 0.04a

N2 0.32a 0.04a 0.11a 0.16a 4.2a 4.1ab 4.0a 4.1a 0.07a 0.01a 0.03a 0.04a

N3 0.04b 0.02a 0.04a 0.03a 4.3a 4.3a 3.8a 4.1a 0.01b 0.01a 0.02a 0.01a

N4 0.04b 0.06a 0.07a 0.06a 4.2a 4.4a 3.8a 4.1a 0.01b 0.02a 0.02a 0.02a

N5 0.21ab 0.04a 0.18a 0.14a 5.2a 4.6a 4.1a 4.7a 0.05ab 0.01a 0.03a 0.03a

Av 0.14A 0.04A 0.08A 4.1A 4.1A 3.7A 0.04A 0.01A 0.03A

2005

N1 0.36b 0.22a 0.15a 0.24b 2.9b 3.3c 2.9c 3.0c 0.10b 0.09a 0.07a 0.09a

N2 0.92b 0.26a 0.45a 0.54ab 5.0a 3.8bc 4.4ab 4.4ab 0.18ab 0.08a 0.10a 0.12a

N3 0.36b 0.26a 0.04a 0.22b 4.6a 4.8ab 3.4bc 4.3b 0.13ab 0.06a 0.01a 0.07a

N4 1.79a 0.71a 0.10a 0.86a 5.2a 5.0a 4.7a 5.0a 0.24a 0.10a 0.03a 0.12a

Av 0.86A 0.36B 0.18B 4.4A 4.2A 3.9A 0.16A 0.08B 0.05B

Data in columns followed by the same a minor letter are not significantly different (P = 0.05). Data

in rows followed by the same capital letter are not significantly different (P = 0.05).

6.3.6. Conclusion


Page | 79

In the long-term aerobic rice field experiment at the IRRI farm, yield of variety Apo declined under

aerobic field conditions relative to flooded conditions in the first 10 season of continuous cropping.

Part of this relative yield decline can be attributed to yield increase under flooded conditions after

the third season, but part is caused by absolute yield decrease under aerobic conditions. The yield

in the same year in a “new” aerobic field (after continuous flooded cropping) was about 2.5 t ha-1

higher than in a seventh-season continuous aerobic field. Compared with the flooded fields,

nematodes of the Meloidogyne graminicola species were much higher in the aerobic fields, but no

correlation with yield was established. The typical patchiness related to the occurrence of

nematodes was not observed.

Yield decline could be reversed by crop rotations, fallowing, and flooding. Flooding for

three consecutive seasons and fallowing for two consecutive seasons were equally effective in

restoring aerobic rice yields. Cropping with upland crops (maize, sweet potato, and soybean) for

two consecutive seasons was more effective than fallowing for two consecutive seasons.

Pot and micro-plot (within the field experiment) experiments showed that yield decline

could not be reversed by the application of micro-nutrients, P, or K. However, crop growth was

consistently improved by the application of N fertilizer in the form of urea or Ammonium sulfate.

Ammonium sulfate, however, was much more effective than urea. The reasons for improved plant

growth with additional N application are not yet understood. It may have to do with changes in soil

pH and subsequent changes in nutrient availability. It is also possible, however, that biotic stresses

(such as nematodes) limit the crop’s ability to take up nutrients and that extra N application may

overcome such a limitation.

Dry season aerobic rice yields increased dramatically after 10 seasons of continuous

cropping. Yields in the 11th and 13th season were 1.4 t ha-1 and 2.1 t ha-1 higher, respectively, than

in the 9th continuous season. This suggests some self-regenerating mechanism in the system but

needs more study to confirm. The increase in aerobic rice yields after 10 seasons was not obvious

in the wet season with lower yield levels than in the dry season.

We found preliminary evidence (data not shown; see Lixiao Nie et al., in prep 2) of

genotypic variation to “soil sickness” associated with continuous aerobic rice cropping. This needs

to be further explored in follow-up studies to see if germplasm improvement can confer some

degree of tolerance to “soil sickness”.

The geographic extent of possible yield decline under continuous cropping or of immediate yield

“collapse” in Asia is not known. The IRRI experiment is the only long-term aerobic rice experiment

in existence. More long-term experiments are needed to determine the extent of the problem.

Beside a few sites in the Philippines, we have not encountered any cases of immediate yield

“collapse” in China or India. Extremely low yields have also been reported in some of our trials in

Laos and Thailand, but these were most likely caused by extreme droughts. The occurrence of

nematodes, however, may be more widespread than assumed so far in rice fields that are not

permanently flooded (Janice Thies, personal communication, from an inventory in S Asia). It is


Page | 80

proposed to conduct diagnostic surveys of “soil health” in current and potential target areas for

aerobic rice.

6.4. Objective 4: Crop establishment

6.4.1. Method

We focused our experiments on seed rate and row spacing. We analyzed and synthesized field

experiments we performed in 2004, and performed new experiments during the STAR project in

2005-2006. In most experiments, the “crop establishment” factor was part of a multi-factor

experiment involving different N regimes or irrigation water regimes. Table 6.4.1 gives an

overview of the location of the experiments.

Table 6.4.1. Sites of crop establishment (seed rate, row spacing) experiments per country

Activity/

Country

Philippines China India Laos Thai

Seed rate Beijing:

Xibeiwang, Shanzhuang

WTC-IARI

station, Delhi

Row spacing San Ildefonso,

Dapdap, Munoz

China. Two field experiments were conducted in 2004 at Xibeiwang village (39°95′ N, 116°4′ E; 43

m asl), Beijing. In the first, three fertilizer N applications (120 kg N ha-1 single application as

coated urea; 120 kg N ha-1 three-split application as regular urea; and 0 N control) were combined

with three seed rates (60, 90, and 120 kg ha-1). Aerobic rice HD297 was dry sown at a depth of 3

cm with a row spacing of 27.5 cm. In the second experiment, seed rates of 60 and 120 kg ha-1

were combined with row spacings of 27.5 and 33 cm. Fertilizer N application was 120 kg N ha-1 in

three-split application using regular urea. The fields of both experiments had 61% sand, 28% silt,

and 11% clay (groundwater deeper than 20 m).

In 2006, we introduced two seed rates in the water x nitrogen experiment at Shangzhuang

near Beijing: 135 (D2) and 67.5 (D1) kg ha-1 (see paragraph “Water and nutrient dynamics” above

for full experiment explanation). The row spacing was kept unchanged at 30 cm; aerobic rice

variety HD297 was used.

India. A water x variety x seed rate experiment was conducted at the WTC-IARI station, Delhi, in

2006-2007. The seed rates were 20, 30, 40, and 80 kg ha-1 (Note that 80 kg ha-1 was used in the

irrigation water experiments reported above). The same 3 irrigation water treatments were used


Page | 81

as in the water experiments: irrigation water applied (through flash flooding) when soil water

tension at 20 cm depth reached “0 kPa” which means keeping the soil close to saturation (I0),

when soil water tension reached 20 kPa (I20), and when soil water tension reached 40 kPa (I40).

The varieties used were Pusa Sugandh 3 and Pusa Rice Hybrid 10. Seeds were dry sown in rows

spaced 25 cm apart.

Philippines. The effect of row spacing was studied in N fertilizer x row spacing interaction

experiments at San Ildefonso (Bulacan), Munoz-PhilRice station, and Dapdap village (Tarlac), in

2004-2005. At all sites, rice variety Apo was dry seeded, and the row spacings were 25, 30, and

35 cm. More experimental details are presented in the paragraph “Water and nutrient dynamics”.

6.4.2. Results China

Yields of HD297 in the experiment on seed rate x fertilizer N at Xibeiwang in 2004 varied from 3.2

to 3.9 t ha-1 (Table 6.4.2). There were no significant differences in yield among neither the seed

rate treatments nor the fertilizer N treatments. Like in the other fertilizer N experiments near

Beijing (see paragraph “Water and nutrient dynamics” above), there was no effect of N application

over the 0 N control. In the second experiment, on seed rate x row spacing, yields were 3.5 to 3.9

t ha-1 (Table 6.4.3). Yields were not affected by seed rate nor by row spacing.

Table 6.4.2. Biomass, grain yield, and harvest index of aerobic rice HD297 grown under different

fertilization and seeding rate treatments, Xibeiwang, Beijing, 2004.

Fertilizer Seeding

Rate (kg ha–2)

Aboveground

Biomass (g m–2)

Grain yield

(g m–2)

Harvest

Index (%)

Coated urea 60 1090 bcd 397 a 39 a

90 1206 abc 368 ab 31 abc

120 1262 a 344 ab 27 c

split N 60 1112 abcd 377 ab 34 abc

90 1238 ab 361 ab 33 abc

120 1052 cd 316 b 31 abc

CK 60 1086 bcd 392 a 36 abc

90 1010 d 349 ab 37 ab

120 1149 abcd 336 ab 29 abc

Different lowercase letters within a column for the same year indicate significant differences at a

P<0.05 level.


Page | 82

Table 6.4.3. Yield and yield components of aerobic rice HD297 grown at different seeding rate and

row spacing treatments, Xibeiwang, Beijing, 2004.

Seeding

rate (kg ha–2)

Row

Spacing (cm)

Yield

(g m–2)

Panicle

m–2

Spikelet

panicle–1

% grain

filling

1000-grain

weight (g)

27.5 376.6 a 266.9 b 80.6 ab 78.6 a 28.4 ab 60

33 396.1 a 212.0 c 86.9 a 71.6 bc 29.0 a

27.5 361.0 a 295.3 a 74.9 b 66.5 c 27.4 b 90

33 349.9 a 226.8 c 76.9 b 68.4 bc 28.5 a

Different lowercase letters within a column indicate significant differences at a P<0.05 level

At Shangzhuang, reducing the seed rate from 135 to 67.5 (D1) kg ha-1, did not significantly affect

yield except in the W1N0 treatment (highest irrigation water application, 0 N application) where

yield was reduced with about 1 t ha-1 (Figure 6.4.1). Among the yield components, the number of

panicles m-2 and the number of spikelets panicle-1 were most affected, though in opposite

directions so that the reduction in the first was compensated by an increase in the second (Figures

6.4.2a,b).

Figure 6.4.1. Yield of HD297 at two seed rates (D1, D2) in four water x nitrogen treatments,

Shangzhuang, near Beijing, 2006. Treatment abbreviations are explained in the methods sections

of “Water x nutrient interaction” paragraph.

0

1000

2000

3000

4000

5000

6000

W1N0 W3N0 W1N2 W3N2

-50

-40

-30

-20

-10

0

10

20

30

40

50D1 D2Final grain yield (kg ha-1)

ba

a aa a

aa

Difference(%)


Page | 83

Figure 6.4.2a. Number of panicles m-2 (A) and number of spikelets panicles-1 (B) of HD297 at two

seed rates (D1, D2) in four water x nitrogen treatments, Shangzhuang, near Beijing, 2006.

Treatment abbreviations are explained in the methods sections of “Water x nutrient interaction”

paragraph.

A

B

6.4.3. Results India

In 2006, yields of Pusa Rice Hybrid 10 were 2.9-4.2 t ha-1, whereas those of Pusa Sugandh 3 were

1.5-3.5 t ha-1 (Figure 6.4.3; data of 2007 still being analyzed). The hybrid variety showed little

yield loss with increasing soil water tension, whereas yields of the inbred variety dropped fast with

increasing soil water tension. In the hybrid variety, yields were at par with 80 and 40 ka ha-1 seed

rate, but yields dropped significantly with seed rates less than 40 kg ha-1. In the inbred variety,

yields decreased significantly with decreasing seed rate below 80 ka ha-1.

0

20

40

60

80

100

W1N0 W3N0 W1N2 W3N2

-50-40-30-20-1001020304050

D1 D2Spikelets/panicle

aa a

a

ab a

b

Difference(%)

0

100

200

300

400

500

W1N0 W3N0 W1N2 W3N2

-50-40-30-20-1001020304050

D1 D2Panicles/m2

a

aa

b b

a

aa

Difference(%)


Page | 84

Figure 6.4.3. Yield of two rice varieties under three irrigation treatments (I0, I20, I40)

and four seed rates, WTC-IARI station, Delhi, 2006.

6.4.4. Results Philippines

Information on groundwater depth and soil water tensions during the experiments, and the effect

of fertilizer N on yield, was presented in the paragraph “Water and nutrient dynamics”. More

detailed results are presented by Lampayan et al. (in prep). In both experiments, there was no

significant effect of row spacing on yield (Table 6.4.4, Table 6.4.5).

0

100

200

300

400

500

I0 I20 I40 I0 I20 I40

Pad

dy y

ield

(g

m-2

)

80

40

30

20

Pusa Rice Hybrid 10 Pusa Sugandh 3

0

100

200

300

400

500

I0 I20 I40 I0 I20 I40

Pad

dy y

ield

(g

m-2

)

80

40

30

20

80

40

30

20

Pusa Rice Hybrid 10 Pusa Sugandh 3


Page | 85

Table 6.4.4. Grain yield and total biomass (averaged over N rates) of Apo in different row spacings

at San Ildefonso and Dapdap, 2004 and 2005 wet seasons.

Site and year Row spacing Yield (t ha-1) Total biomass

at harvest (t ha-1)

San Ildefonso, 2004 RS25 3.76 a 14.22 b

RS30 3.72 a 12.74 a

RS35 3.69 a 12.04 a

San Ildefonso, 2005 RS25 3.85 ab 10.12 b

RS30 3.91 b 10.10 b

RS35 3.61 a 8.94 a

Dapdap, 2004 RS25 3.65 a 10.58 b

RS30 3.65 a 9.78 a

RS35 3.69 a 9.30 a

Dapdap, 2005 RS25 3.20 a 7.45 b

RS30 3.17 a 7.29 ab

RS35 3.15 a 6.48 a

Means followed by a different letter (within same site and season) are significantly different at the

5% level.


Page | 86

Table 6.4.5. Effects of row spacings on grain yield (averaged over N splits) and yield components

of Apo cultivar in PhilRice during 2004 wet season and 2005 dry season.

Year

Row spacing


at harvest (t ha-1)

2004 RS25 3.49 a 10.2 b

RS30 3.40 a 9.3 a

RS35 3.58 a 9.1 a

2005 RS25 5.79 a 11.6 a

RS30 5.82 a 11.6 a

RS35 5.84 a 11.4 a

Means followed by a different letter (within same site and season) are significantly different at the

5% level.

6.4.5. Conclusion

Direct dry-seeded aerobic rice (varieties Apo in the Philippines, and HD297 in China) does not

seem to be very responsive to row spacing and seed rate (within the limits tested). Both in the

Chinese and the Philippine experiments, row spacing between 25 and 35 cm gave statistically the

same yields. In the Chinese experiments, yields were the same with differences in seeding rate in

the 60-135 kg ha-1 range. In the India experiment, the better performing variety Pusa Rice Hybrid

10 had statistically same yields with seed rates of 40 and 80 ka ha-1, but below 40 kg ha-1, yields

declined fast. In practice, the “unresponsiveness” of aerobic rice to seed rate and row spacing

means that farmers are rather flexible in choosing their own rates and spacings (within the limits

studied). A close row spacing and high seed rate may increase the competitiveness of aerobic rice

to weeds but may increase seed costs. A wider row spacing may facilitate interrow cultivation,

such as mechanical weeding, and reduce seed costs.

6.5. Objective 5: Target domain

Aerobic rice is an option to farmers, in either irrigated or rainfed areas, who would like to grow rice

but where water availability at the farm level is too low, or where water is too expensive, to grow

flooded rice. There are many water-saving technologies for rice, and the most suitable or

“attractive” technology depends on the type and level of water scarcity, soil properties to hold

water, availability of suitable rice varieties, the irrigation infrastructure (if present), and the socio-

economics of their production environment (Bouman et al., 2006, 2007). In this paragraph, we

review these factors to describe the target domain for aerobic rice, with the exception of “irrigation

infrastructure” as that was beyond the scope of our CPWF project. The methods we used were field

experiments, simulation modeling, GIS, and household surveys, with a focus on China (Table


Page | 87

6.5.1). We also report briefly on the initial results of extrapolation domain analysis that was part

of the CPWF’s Impact project.

Table 6.5.1. Sites of target domain analysis per country

Activity/country Phil China India Laos Thai

Field experiment Changping, Beijing

Variety zoning Whole country

Modeling, GIS Beijing, Tianjin,

Shandong, Hebei,

Henan

Household survey Kaifeng, Fengtai,

Yingshang, Beijing

Extrapolation

domain

Whole

country

Whole country Whole

country

Whole

country

Whole

country

6.5.1. Water availability and soil type

Figure 6.5.1 presents a gradient in water availability to grow a crop at the field level, and some

appropriate technologies to grow rice. Going from right to left along the water-availability axis,

water gets increasingly scarce and yields decline. On the far right-hand side of the water axis,

water is amply available and farmers can practice continuous flooding and obtain the highest

yields. With decreasing water availability, appropriate technologies are saturated soil culture or

alternate wetting and drying. With further decreasing water availability, trying to keep growing

lowland rice (aiming for saturated soil conditions and at least partial flooding) is not be possible

anymore and aerobic rice becomes attractive. On the far left-hand side, water is extremely short,

such as in rainfed uplands, and yields are very low. Here, sturdy and drought-resistant upland

varieties and upland cropping systems become appropriate.


Page | 88

Figure 6.5.1. Schematic presentation of appropriate water-saving technologies along a relative

water availability axis. AWD = alternate wetting and drying, SSC = saturated soil culture, FC =

field capacity, S = saturation point, ∆Y = change in yield. Source: Bouman et al. (2007).

How much less water is used under aerobic conditions than under flooded conditions depends

mostly on the seepage and percolation (SP) losses under flooded conditions and on the deep

percolation losses of irrigation water under aerobic conditions. Typical SP rates of flooded rice

fields vary from as little as 1 mm d-1 to more than 25 mm d-1 (Bouman and Tuong, 2001). Under

aerobic conditions, the amount of deep percolation depends on the combination of soil water

holding capacity and method of irrigation, and is reflected in the irrigation application efficiency

(EA). With a precise dosage and timing of irrigation in relation to crop transpiration and soil water

holding capacity, the EA in flash-flood irrigation can be up to 60% (Doorenbos and Pruit, 1984). If

furrow irrigation (or raised beds) is used, the EA can go up to 70%, and with sprinkler irrigation up

to 80% or more. Assuming an average growth duration of 100 days, and mean ET values for rice,

we can roughly calculate the “break-even” point for SP rates in flooded fields that would result in

similar water requirements in aerobic fields with different irrigation methods (Table 6.5.2). When

the SP rate in flooded rice is 3.5 mm d-1 or higher, aerobic systems with flash-flood irrigation will

require less water, and if the SP rate is 0.5 mm d-1 or lower, only aerobic systems with sprinkler

irrigation require less water. When aerobic rice systems are direct (dry) seeded, as is the typical

target technology, an additional amount of water input can be saved by foregoing the wet land

preparation.

0

5

10

Furrow, Flush,

Sprinkler Irrigation SSC

Soil Management,

Reduced water depth

∆Y = f (variety, management)

Technologies to reduce water input

AWD

Lowland rice

system

Aerobic rice

system

Traditional

upland rice

system

Water availability High Low

Soil Condition

FC S

Aerobic Flooded

Yield (t ha-1)


Page | 89

Table 6.5.2. Comparison of water use in a hypothetical aerobic rice crop with that of lowland rice

on different soils types characterized by their seepage (S) and percolation (P) rates. Source:

Bouman et al., 2005.

Water flow process Aerobic rice (mm) Lowland rice (mm)

Lowland soil SP rate - - 1 mm d−1 5 mm d−1 15 mm d−1

Irrigation efficiency 85% 60% - - -

Evaporation 100 100 200 200 200

Transpiration 400 400 400 400 400

Seepage and percolation − − 100 500 1,500

Irrigation inefficiency loss 90 335 − − −

Total 590 835 700 1,100 2,100

An example of the cross-over point in terms of water availability where aerobic rice gives higher

yields than flooded lowland rice is given in Figure 6.5.2 for our field experiments at Changping,

Beijing, (see paragraph “Water and nutrient dynamics”). Two aerobic rice varieties (HD297 and

HD502) and one lowland rice variety (JD305) were grown under flooded conditions and under

aerobic soil conditions with different amounts of total water input (irrigation and rainfall). Under

flooded conditions with 1300-1400 mm water input at the right-hand side of the horizontal (water)

axis, the lowland variety JD305 gave highest yields of 8-9 t ha-1. The yield of JD305, however,

quickly declined with increasing water shortage and aerobic soil conditions. With less than 900 mm

water input, and under aerobic soil conditions, the aerobic rice varieties HD297 and HD502

outperformed the lowland variety. To further explore the cross-over point where aerobic rice

systems are more suitable than lowland rice systems, comparisons of aerobic rice with systems

such as saturated soil culture or alternate wetting and drying should be made. However, we can

conclude that, in this environment, water availability for aerobic rice should be around 400-900

mm during the growing season.


Page | 90

Figure 6.5.2. Yield of aerobic rice varieties (black diamonds) and a lowland variety (open

diamonds) under flooded and aerobic soil conditions, Changping, 2001-2004.

0

1

2

3

4

5

6

7

8

9

10

200 400 600 800 1000 1200 1400

Yield (t ha-1)

Water input (mm)

aerobic soil flooded soil

6.5.2. Target domain varieties

The availability of suitable varieties is an essential component of identifying target domains. Even

when sufficient water and the right soil types are present (see paragraph “Water availability and

soil type”), varieties need to be available that fit local climatic conditions and local cropping

patterns. Important factors are yield potential, taste, crop growth duration to fit within the growing

season or certain crop rotations, and resistance to low or high temperatures. In the Yellow River

Basin in China, farmers may be able to grow just one summer crop in the northern part, but may

have a double cropping system of a winter crop (wheat, barley, rapeseed) and a summer crop in

the southern part. In the north, aerobic rice varieties with a long duration are needed to maximize

the length of the summer season, whereas in the south, a short duration crop is needed to allow

for timely establishment and harvest of the winter crop. Tolerance to low temperature in early

spring is important in the north, whereas tolerance to heat during flowering may be important in

both the north and the south.

Since the breeding of aerobic rice varieties has been ongoing since (at least) the early

nineteen eighties, many varieties developed by a number of universities and institutes are

available. In our project, we used expert knowledge by breeders to develop a target domain map

for Han Dao varieties produced by China Agricultural University (Figure 6.5.3). Similar maps can

be produced for other varieties.


Page | 91

Figure 6.5.3. Target domain and estimated distribution of aerobic rice varieties of the Han Dao

series in China.

6.5.3. Mapping yield and water requirements

We used a combination of simulation modeling with ORYZA2000 and GIS to produce maps that

show yield potential and water requirements in part of the Yellow River Basin and the North China

Plain: Beijing, Tianjin, Shandong, Hebei, and Henan provinces. First, a crop data file was created

based on the calibrations for HD297 using experimental data collected at Kaifeng (Feng et al.,

2007) and Changping, Beijing (Xue et al. in prep; see paragraph “Simulation modeling”). Next,

this data file was adapted to northern and southern climatic regions within the YRB/NCP by

changing the crop growth durations to fit the cropping seasons and cropping patterns. Thus, two

“model aerobic rice crops” were created that were both based on HD297 but had different

durations. Next, we collected daily weather data of 50 years between 1951 and 2000, from 69

stations in the 5 provinces of our study. We compiled soil information for the area covered by the

representative weather stations. The required soil hydrological properties to run the SAWAH water

balance model of ORYA2000 were either directly obtained from published data or estimated using

pedotransfer functions on available secondary data (texture, soil organic matter content, bulk

density). Next, ORYZA2000 was run for each year of each weather station using the representative

Handao 65，，，，271

Handao 297，，，，9

Handao 277，，，， 502

Handao 175，，，，8, 9


Page | 92

soil data. We run two scenarios: a potential production situation and a rainfed production situation.

In the potential situation, the soil water content in the root zone was kept continuously at field

capacity, and we computed “potential aerobic rice yield” and crop water requirements by seasonal

evapotranspiration. The difference between crop water requirements and seasonal cumulative

rainfall was interpreted as net irrigation demand (or water deficit). In the rainfed scenario, we

computed rainfed yields with a groundwater table at 1.9 m depth. We computed the 50-year

averages of all simulated variables at each station, and used GIS interpolation techniques to

extrapolate the station values to their surroundings (based on the approach of “simulate first,

average and interpolate after”). So far, we did not formally include topographic features such as

mountains and nearness to the ocean in our interpolation, but selected the interpolation technique

that produced maps that visually corresponded best with such features. The results are shown in

Figures Figures 6.5.3a-e.

Potential yield in the central part of the YRB are 6-8 t ha-1, which is 0-2 t ha-1 higher than the

maximum yield recorded in any of our field experiments. However, in our field experiments, we

never maintained soil water conditions at field capacity. The potential yields merely indicate the

level that could be reached with no stress or water shortages whatsoever. Under farmer

conditions, more realistic “attainable yields” (Van Ittersum and Rabbinge, 1997) are defined which

usually approach 80% of the yield potential, which would translate into 4.8-6.4 t ha-1 in the central

YRB. Rainfed yields in the central YRB are also 6-8 t ha-1, indicating that aerobic rice could

potentially be grown without irrigation. The categories in our yield maps mask the fact that the

rainfed yield are in the lower ranges of each category compared with the potential yields. Most of

the area has an average seasonal rainfall of 450-510 mm and a water deficit of only 0-220 mm,

which had little impact of yields. It should be noted, however, that in our simulations, we started

crop growth with a soil profile down to 1.9 m at field capacity (contributing to crop water

requirements by uprise; Bouman et al., 2007), which in reality may not be the case if a winter

crop is grown (that has exhausted the soil from water). On average, we can conclude that aerobic

rice can be grown under conditions of purely rainfed to about 220 mm of supplementary irrigation

(which would practically translate into 3 gifts of 70 mm). This compares well with actual practices

of farmers who grow aerobic rice (see below). In dry years, of course, irrigation requirements are

higher than in our simulated 50-year average values.


Page | 93

Figure 6.5.3a. Potential aerobic rice yield (t ha-1).

Figure 6.5.3b. Average rainfed aerobic rice yield (t ha-1) with groundwater at 1.9 m.


Page | 94

Figure 6.5.3c. Crop water requirement (mm) by evapotranspiration

Figure 6.5.3d. Seasonal rainfall (mm)

Figure 6.5.3e. Crop water deficit: crop water requirements minus rainfall (mm)


Page | 95

6.5.4. Farmers’ practices

We used data collected in farmers’ fields in Kaifeng before our CPWF project started to complete

the supplement the theoretical analysis of target domain presented above (more details by

Bouman et al., 2007). In 2002 and 2003, a number of farmers grew aerobic rice and we monitored

their yield, water input, and groundwater table depth throughout the growing season. Yield was

obtained by interview as whole field yield and includes post-harvest losses from harvest to delivery

to the miller as air-dry rough rice. All farmers used only shallow tubewells for irrigation and

irrigation water input was computed from measured time of pump operation times the calibrated

flow rate of the pump. Rainfall was taken from Hubei experiment station. In 2005, in our CPWF

project, we interviewed farmers who had adopted aerobic rice about their yields and harvested

areas.

The groundwater table depths observed in farmers’ fields are given in Figure 6.5.4 and

demonstrate the variability from shallow to deep (more than 2 m) that can occur in the same area.

The yield and water use are given in Table 6.5.3. Yields were lower in 2003 than in 2002

because of heavy rains during flowering in 2003, which resulted in increased spikelet sterility (the

same rainfall that caused the groundwater tables to sharply rise, Figure 6.5.4). Most farmers

applied three irrigations in 2002, and, because of more rainfall, only two in 2003. Irrigations were

usually given to promote germination after sowing, and between tillering and flowering when there

was not enough rainfall. There was no relationship between yield and irrigation water input, nor

between yield and groundwater depth.

The yields obtained by farmers in 2005 are given in Figure 6.5.5. Out of 36, five farmers

harvested no yield at all (fields were abandoned), and six farmers had yields between 5000 and

5500 kg ha-1. Excluding the abandoned fields, the average yield was 3380 kg ha-1 and including

the abandoned fields, it was 2900 kg ha-1.


Page | 96

Figure 6.5.4. Measured groundwater depth in aerobic rice fields of farmer A (♦), B (◊), C (♦), G

(∆), and H (*) in 2002 (a), and of farmer V (♦), W (◊), X (♦), and Y (∆) in 2003. Groundwater

depths at fields of other farmers listed in Table 6.5.3 were not measured.

-250

-200

-150

-100

-50

0

50

8-Jun 28-Jun 18-Jul 7-Aug 27-Aug 16-Sep 6-Oct 26-Oct

aGroundwater depth (cm)

Date

-250

-200

-150

-100

-50

0

50

8-Jun 28-Jun 18-Jul 7-Aug 27-Aug 16-Sep 6-Oct 26-Oct

bGroundwater depth (cm)

Date

Figure 6.5.5. Frequency distribution of aerobic rice yields by 36 farmers in 2005.

0

5

10

15

20

25

0-0.5 0.5-1 1-1.5 1.5-2 2-2.5 2.5-3 3-3.5 3.5-4 4-4.5 4.5-5 5-5.5 5.5-6

YYield (t ha-1)

Frequency (%)


Page | 97

Table 6.5.3. Performance of aerobic rice farmers in terms of yield and water use, Kaifeng, 2002-

2003.

Farmer label A B C D E F G Mean

2002

Field size (ha) 0.17 0.11 0.13 0.07 0.07 0.04 0.21 0.12

Yield (kg ha-1) 3800 4400 3800 5100 5500 4700 3400 4300

Irrigation (mm) 225 225 80 231 230 300 225 217

Rainfall (mm) 337 337 337 337 337 337 337 337

Total water input (mm) 562 562 417 568 566 637 562 553

2003

Farmer U V W X Y Z Mean

Field size (ha) 0.10 0.16 0.06 0.11 0.11 0.13 0.11

Yield (kg ha-1) 1200 3825 2400 3750 3608 3038 2970

Irrigation (mm) 156 159 145 169 146 162 156

Rainfall (mm) 674 674 674 674 674 674 674

Total water input (mm) 830 833 818 842 820 836 830

We also synthesized initial socio-economic survey data among farmers in 2002-2003 before our

CPWF project started, and did a larger survey in 2005. Beside aerobic rice, the survey included

other crops such as lowland rice and upland crops (maize, soybean, cotton, etc). The purpose was

a general comparison among alternative crops available to farmers. In 2001-2003, we surveyed a

small number of farmers in different sites (Table 6.5.4, Table 6.5.5). In total, 24 aerobic rice fields

were included, 20 lowland rice fields, and 21 upland cropped fields. In 2005, we surveyed about 60

households in Kaifeng, Henan, and Fengtai, Anhui, counties. The surveys were accompanied by

focus-group discussions to get farmers’ opinions on adoption of aerobic rice.


Page | 98

Table 6.5.4. Location and year of aerobic rice surveys, China, 2001-2003.

County Village Year

Beijing Changle, Hanjiachuan 2001

Yingshang Gaogu, Tulou 2001

Fengtai Guanzhuang, Dacheng, Zhaixi 2002-2003

Kaifeng Mengtang, Panlou, Shangzai, Shunji,

Xilong

2002-2003

Table 6.5.5. Number of crops (fields/farms) included in the aerobic rice surveys, China, 2001-

2003.

Year Crop Number of fields

2001 Aerobic rice 4

Lowland rice 4

2002 Aerobic rice 11

Lowland rice 11

Cotton 3

Maize 3

Sesame 2

Soybean 1

2003 Aerobic rice 9

Lowland rice 5

Cotton 4

Maize 6

Soybean 2

In 2001-2003, returns over paid-out costs to aerobic rice varied from 234 to 591 $ ha-1 (Table

6.5.6). Usually, this was lower than for upland crops. Labor use was about the same level as for

maize and soybean, but much lower than for cotton.


Page | 99

Table 6.5.6. Results of aerobic rice surveys, China, 2001-2003.

In 2005, net returns over costs (including own labor use) to aerobic rice was 326 $ ha-1 (Table

6.5.7). This was higher than for maize and soybean, a bit lower than for lowland rice, and much

lower than for the cash crops peanut and cotton. Labor use was higher than for soybean, about the

same level as for maize and peanut, and lower than for lowland rice.

Overall, we conclude that aerobic rice can be a financially attractive crop, but that profits need to

increase to make it more competitive. In focus-group discussions, farmers mentioned that 6 t ha-1

is an ideal yield target, while most of the surveyed farmers have yields of 3-4 t ha-1. Aerobic rice

uses much less irrigation water (156-217 mm) than lowland rice (1300-1500 mm; Wang Huaqi et

al., 2002; Feng et al., 2007), and can well be grown in areas where water shortage excludes the

growing of lowland rice. When sufficient water is available, farmers should grow lowland rice

instead of aerobic rice as yields and profits are higher. The range of irrigation water application by

farmers supports the results of our modeling and GIS study (see above). Farmers mentioned the

following reasons for adopting aerobic rice: they want to grow rice when water is insufficient (for

lowland rice), ease of establishment, less labor use (compared with lowland rice), good eating

quality. Negative views included: low yields, difficult to control weeds, and insufficient extension

support. One special positive characteristic of aerobic rice is that it can stand flooding in situations

where all other upland crops would suffer or even get completely wiped out. In the peak towards

end of summer time, the Yellow River and its tributaries often overflow and cause flooding for a

few days to weeks. Under these conditions, aerobic rice still survives and can produce a

harvestable yield.


Page | 100

Table 6.5.7. Results of aerobic rice surveys, China, 2005.

6.5.5. Extrapolation domains

We collaborated with the Impact Project of the CPWF to develop extrapolation domains to explore

the impact potential of aerobic rice. Although this work is still in progress, we report summary

results to date (taken from Rubiano et al., 2007). Potential extrapolation domain areas for aerobic

rice were calculated using Homologue and Weights of Evidence modelling that look for similar

agro-ecological and socio-economic conditions to those found in project pilot sites. The analysis

shows that the highest probability areas are all in Asia (Figure 6.5.6). In India, the extrapolation

domain is largely centred on the rice-wheat systems in the Indo-Gangetic basin. In Thailand and

Burma the areas are centred on rainfed lowland areas. The analysis found large areas that are

suitable climatically in Africa in Zimbabwe, Mozambique, Madagascar, Burkina Faso and Nigeria,

and in Latin America, in Brazil, Bolivia and Venezuela.

The pan-tropical perspective of the extrapolation domain analysis is both its strength and

weakness. The extrapolation domain analysis is restricted to using publicly available pan-tropical

databases that necessarily restrict the socio-economic variables that can be modelled. The

extrapolation domain areas and the quantification of potential impacts presented in this paper are

important not in absolute terms, but in the trends they show. Extrapolation domain and scenario

analysis are useful not for predicting what will happen, but rather exploring what could happen.

Comparative profitability of rice production and other crops in Kaifeng Henan and Fengtai Anhui, China 2005

CROPS Aerobic Rice

Lowland

Rice CORN SOYBEAN PEANUT COTTON

n 59 16 101 43 7 10

Yield 4.18 5.76 4.76 0.25 0.30 1.09

Gross return 967.19 1,316.14 708.98 423.45 1,315.44 1,249.33

Cost of Production

Material inputs 356.61 413.74 182.40 177.95 317.81 202.25

Fertilizer cost 137.02 178.13 104.61 65.19 111.88 52.36

Insecticide cost 18.02 27.09 11.49 7.01 6.45 28.40

Herbicide cost 19.77 5.76 - - - -

Seed cost 88.02 63.20 34.23 74.52 161.40 107.87

Power cost 54.67 75.96 19.12 28.08 32.20 13.19

Irrigation Cost 16.31 34.46 - - - -

Other Cost 22.79 29.14 12.95 3.15 5.86 0.43

Labor cost 284.86 460.21 225.43 111.98 282.43 395.35

Hired labor 2.89 48.42 - - - -

Imputed family labor cost 281.97 411.78 225.43 111.98 282.43 395.35

Total Paid out cost 359.50 462.17 182.40 177.95 317.81 202.25

Total cost 641.47 873.95 407.83 289.93 600.23 597.60

Return over paid out costs 607.69 853.97 526.58 245.50 997.64 1,047.09

Net return 325.73 442.18 301.15 133.52 715.21 651.73


Page | 101

Figure 6.5.6. Extrapolation domains for STAR aerobic rice in tropics in Asia

6.5.6. Conclusion

Aerobic rice holds promise for farmers in water-short irrigated or rainfed environments where

water availability at the farm level is too low, or where water is too expensive, to grow flooded

lowland rice. When the percolation losses in flooded rice are 3.5 mm d-1 or higher, aerobic systems

with flash-flood irrigation will require less water, and if the percolation losses are 0.5 mm d-1 or

lower, only aerobic systems with sprinkler irrigation require less water. In northern China, the

target areas are where water availability (rainfall with or without supplementary irrigation) is 400-

900 mm during the cropping season. Suitable varieties here are HD297 and HD502 (among

others). In the central part of the Yellow River Basin (Kaifeng area), and in most of the North

China Plain, attainable yields of 5-6 t ha-1 are possible with 0-220 mm irrigation application (in

average rainfall years; groundwater 2 m deep or less). In these regions, farmers who currently try

out aerobic rice attain usually 3-4 t ha-1, with 150-220 mm supplementary irrigation (though some

farmers get up to 5.5 t ha-1). With these yields, financial returns to aerobic rice can be more or

less than to upland crops (maize, soybean), depending on relative market prices (that fluctuate

among years). Yields need to go up to 6 t ha-1 to make aerobic rice more competitive. Reasons for

adoption are: having own rice on the farm, ease of establishment and less labor needs (compared

with lowland rice), good eating quality. Negative views are: low yields, difficult to control weeds,

insufficient extension support, difficult to market. Aerobic rice is unique in its characteristics to

withstand both flooding and dry soil conditions, which make it an ideal crop for areas prone to

surface flooding where other crops would suffer or fail. Extrapolation domain construction and


Page | 102

scenario analysis suggest that aerobic rice can have large impacts in India and countries in the

lower parts of the Mekong Basin.


Page | 103

7. INTERNATIONAL PUBLIC GOODS

The main international public good (IPG) produced by the project is the concept of aerobic rice: a

new rice production system in which specially developed, input-response rice varieties with

“aerobic adaptation” are grown in well-drained, nonpuddled, and nonsaturated soils without

ponded water. Aerobic rice is aimed at water-short irrigated or rainfed environments where water

availability is insufficient to grow flood-based lowland rice. Although our project did not actually

breed new rice varieties suitable for aerobic production systems, it did identify suitable and

released varieties

• Northern China: HD277, HD297, HD502

• India: Pusa Rice Hybrid 10, Proagro6111 (hybrid), Pusa834

• Philippines: “Apo” (PSBRc9), UPLRI5, PsBRc80

Released varieties are available without restrictions through private seed companies or national

public seed distribution channels (including universities and institutes). For Thailand and Laos, only

suitable genotypes (breeding “lines”) were identified for further breeding (see paragraph “Aerobic

rice varieties identified).

Our project produced initial management options and guidelines with respect to crop

establishment, irrigation, and fertilization. Aerobic rice is basically managed like a wheat or a

maize crop. The usual establishment method is dry direct seeding. Before sowing, the land should

be dry prepared by ploughing and harrowing to obtain a smooth seed bed. Seeds should be dry

seeded at 1-2 cm depth in heavy (clayey) soils and 2-3 cm depth in light-textured (loamy) soils.

Optimum seeding rates still need to be established but are probably in the 70-90 kg ha-1 range. In

experiments so far, row spacings between 25 and 35 cm gave similar yields. The sowing of the

seeds can be done manually (e.g., dibbling the seeds in slits opened by a stick or a tooth harrow)

or using direct seeding machinery. An alternative establishment method is transplanting, where

seedlings are transplanted into wet soil that is kept around saturation for a few days to ease

transplanting shock. Subsequently the fields dry out to field capacity and beyond. This method of

crop establishment can only be done in clay soils with good water-holding capacity. An aerobic rice

crop attaining 4-6 t ha-1 yields obtains many of its required nutrients from the soil. But this

“indigenous supply” of nutrients is typically not sufficient to meet all the nutrient needs, and

fertilizers will need to be applied. Site-specific knowledge about the indigenous nutrients supply is

the starting point for formulating fertilizer recommendations. The site-specific nutrient

management (SSNM) approach (www.irri.org/irric/ssnm) can be used to determine the need for

supplemental nutrients in the form of fertilizers and the optimal management of fertilizers. A

useful tool to assist in the application of nitrogen (N) fertilizer is the Leaf Color Chart (LCC; part of

the SSNM approach). In the absence of trained extension personnel in SSNM and LCC, an amount

of 70-90 kg N ha-1 could be a useful starting point (to be subsequently optimized). Instead of basal

application of the first N split, the first application can best be applied 10-12 days after emergence

to minimize N losses by leaching (the emerging seedling can’t take up N so fast, so it will easily

leach out). Moreover, basal application of N also promotes early weed growth. Second and third

split applications of N may be given around active tillering and panicle initiation, respectively. Dry,


Page | 104

aerobic, soil can reduce the indigenous supply of phosphorus (P), hence the application of fertilizer

P can be more critical for aerobic rice than for conventional flooded lowland rice. On acid soils,

aerobic rice will likely be less prone to zinc deficiency than flooded lowland rice; but on high pH

soils with calcium carbonate, the reverse may be true. If the crop is grown in a dry season, a light

irrigation application (say 30 mm) should be given after sowing to promote emergence.

Subsequent irrigation applications depend on the rainfall pattern, the depth of groundwater, and

on the availability and/or cost of irrigation water. Irrigation can be applied by any means as used

for upland crops: flash flood, furrow, or sprinkler. Rice that is not permanently flooded tends to

have more weed growth and a broader weed spectrum than rice that is permanently flooded. To

control weeds, the use of pre- or post-emergence herbicides is recommended when the weed

pressure is high, plus additional manual or mechanical (inter-row cultivation) weeding in the early

phases of crop growth.

The methods we used were pot experiments, field experiments, simulation modeling, GIS, and

socio-economic household surveys. The crop growth simulation model was evaluated for aerobic

rice variety HD295 and aerobic field conditions, and can be freely obtained from IRRI

(http//www.knowledgebank.irri.org/oryza2000/). Experiment protocols and survey forms used can

be obtained from the project partners. No new research technologies or methodologies were

developed in this project. Experimental data are reported in this report and in national and

international publications.


Page | 105

8. OUTCOMES (PARTNER CHANGE)

Before this CPWF project, IRRI and a few partners in China, India, and the Philippines initiated

research and development activities on aerobic rice in 2001, coordinated through the Water

Workgroup of the Irrigated Rice Research Consortium (IRRC). Until 2004, activities were modest,

operating on small budgets. The CPWF allowed us to increase the scale of our R&D operations

considerably and both scale up and out. The number of partners increased, and we opened new

sites with new partners in Laos and Thailand. The biggest partnership changes occurred in China

and the Philippines. Although the seeds for change were sown from 2001 onward, it was the CPWF

that was instrumental in bringing these changes to fruition.

In China, breeders have been developing aerobic rice varieties since the early nineteen eighties,

and extending their varieties through upland rice networks and channels. Without much research

on the management and performance of these varieties, breeders interacted directly with

extension networks and farmers. Through the CPWF, we were able to add a research dimension,

especially on crop physiology, plant nutrition, agro-hydrology, modeling, and crop management in

general, which was not present before. Initially primarily the domain of breeders, now the aerobic

rice effort at China Agricultural University is led by a multi-disciplinary team from various

departments. The activities of this team, and the status of working in an international partnership,

have attracted the attention of university leaders and increased the support for aerobic rice

research. Members of the team were able to secure additional national funds to support their

aerobic rice research. Moreover, the exposure to the concept of impact-pathways during the

Zengzhou workshop organized by the CPWF in June 2007, increased the awareness of the team

members of “thinking beyond research” (though no tangible results of this change can yet be

reported). Besides working in well-controlled field experiments, some of the team members

initiated on-farm research and participated in socio-economic surveys among farmers in China.

The need for studying impact and adoption has become evident and team members also

participated as resource persons in an externally commissioned impact and adoption study. The

“changes in partner” may be illustrated among a circular Impact-Pathway (Figure 8.1). In 2001,

CAU breeders operated on the left side of the circle, working on extension and farmer-uptake of

aerobic rice through private sector (seed companies) partnership and local government support.

However, there was no research component and questions on how to best manage the new

aerobic rice varieties in a sustainable and profitable way were left unanswered. Mainly through the

CPWF project, a basic and applied research component was added to answer these questions

(arrow 1; started in 2001, but on a small scale). Also, there was little knowledge on the actual

extent of farmer uptake of aerobic rice, reasons for adoption, and impacts of adoption. Through

the CPWF, awareness of the need for insight into these processes led to the first impact

assessments and adoption studies of aerobic rice in northern China (arrow 2). In summary, the

“partner change” was the addition of a scientific research base on crop management and

impact/adoption to the existing breeding-extension network, thereby enriching the impact-

pathway circle.


Page | 106

Figure 8.1. Partnership change (CAU), China 2001-2007.

In the Philippines, the network of partners expanded rapidly with the CPWF. Again, the seeds of

change were sown from 2001 onward, but the CPWF allowed us to scale out in terms of partners.

Originally, two Philippine national partners were included as official project partners: PhilRice, and

the National irrigation Administration in Tarlac. Through the success and visibility of their activities

(they organized and attended many national workshops and trainings), more and more R&D

partners became interested in aerobic rice. The most prominent are listed as project partners

“without CPWF funds”: Bulacan Agricultural State College (BASC), Central Luzon State University

(CLSU), and the National Soil and Water Resources Research and Development Center - Bureau of

Soil and Water Management (BSWM). These partners consider themselves full member of the

aerobic rice project and fully participated in project planning and discussion meetings and

workshops. These partners picked up different components of the Impact-pathway circle according

to their own mandates and interests, ranging from research to teaching, training, and extension.

CLSU, through its Water Resources Management Bureau, introduced aerobic rice in trainings on

water management and embarked on research on sprinkler irrigation of aerobic rice. BASC and

BSWM jointly initiated a range of applied research activities and started extending the aerobic rice

to farmers in the area. Each year, trainings were provided and farmer field schools organized. The

case of BASC-BSWM is illustrated again with the Impact-pathway circle (Figure 8.2). In 2001-

2003, NIA collaborated with IRRI on aerobic rice field experiments in Tarlac. In 2003, one NIA staff

member transferred to BASC and motivated the college to develop an aerobic rice RD&E program

to serve their students and farmers in the province (arrow 1). Applied research on crop

management and variety selection was started in 2004 and continues up to today. Also, farmer

training, demonstrations, and field schools in the neighboring areas were initiated and support

mobilized through Local Government Units (LGU) and village leaders. Within a few years, a rather

complete set of activities along the Impact-pathway circle has been developed at the local scale

Basic research

Applied research

Participatory R&D

Training

Extension

Policy support

Private sector

Farmer uptake

NARES uptake

Impact

Local support

Technologies

Adaptation

2001: 80,000 haAerobic rice ??

CAU/IRRI: experimentsVariety, water 2001-2002Water x N 2003-…Establishment 2004-…

WU, Kaifeng: 2002-2005

Mencheng, 2005-2007Initial surveys 2001-2004

Economic survey60 households, 2005

ZUEL: Adoption survey500 farmers, 2007 (IRRC-CPWF)

1

?2

2001: existing breeders’Networks, CAU key player

Basic research

Applied research

Participatory R&D

Training

Extension

Policy support

Private sector

Farmer uptake

NARES uptake

Impact

Local support

Technologies

Adaptation

Basic research

Applied research

Participatory R&D

Training

Extension

Policy support

Private sector

Farmer uptake

NARES uptake

Impact

Local support

Technologies

Adaptation

2001: 80,000 haAerobic rice ??

CAU/IRRI: experimentsVariety, water 2001-2002Water x N 2003-…Establishment 2004-…

WU, Kaifeng: 2002-2005

Mencheng, 2005-2007Initial surveys 2001-2004

Economic survey60 households, 2005

ZUEL: Adoption survey500 farmers, 2007 (IRRC-CPWF)

11

?2

?2

2001: existing breeders’Networks, CAU key player


Page | 107

(encompassing s few villages, but ambition to scale up and out nationally!). BASC and BSWM were

successful in applying for national R&D grants to finance their activities, and it was their

participation in high-profile programs such as the IRRC and the CPWF that contributed to that

success.

Figure 8.2. Partnership change (BASC-BSWM), Philippines, 2004-2007.

The above examples highlight two “partner changes” of the project All project partners benefited a

lot from working in partnerships (both IRRC and CPWF), and the value of interaction at joint

planning meetings, trainings, and workshops can not be over-stated. Though we did not

systematically document it, we believe that significant “partner changes” along the Impact-

pathway have happened over time. Three key lessons we learned are:

1. Let partners take ownership, and give them freedom to modify and adapt concepts and

practices.

2. Be flexible in partnership arrangements; follow new initiatives developed by new partners;

include new and exciting partners in the project. Unfortunately, this is usually hindered by

project arrangements (such as in the CPWF project contracts) in which partners need to be

identified beforehand with their roles and responsibilities exactly identified at the start.

3. Create opportunities for new partnerships through training.

Basic research

Applied research

Participatory R&D

Training

Extension

Policy support

Private sector

Farmer uptake

NARES uptake

Impact

Local support

Technologies

Adaptation

NIA and IRRI early field experiment on aerobic rice, Tarlac (2003)

1 NIA staff membertransfers to BASC

Experiments on W, N,Establishment, variety,etc, BASC and BSWM2004-….

PVS and farmer demonstrations

Farmer field schools

LGU, frontier farmers

Maybe 100 farmers

2007Adoption study

1

2

?

2005-2007: staff member PhD scholar CLSU

Basic research

Applied research

Participatory R&D

Training

Extension

Policy support

Private sector

Farmer uptake

NARES uptake

Impact

Local support

Technologies

Adaptation

Basic research

Applied research

Participatory R&D

Training

Extension

Policy support

Private sector

Farmer uptake

NARES uptake

Impact

Local support

Technologies

Adaptation

NIA and IRRI early field experiment on aerobic rice, Tarlac (2003)

1 NIA staff membertransfers to BASC

Experiments on W, N,Establishment, variety,etc, BASC and BSWM2004-….

PVS and farmer demonstrations

Farmer field schools

LGU, frontier farmers

Maybe 100 farmers

2007Adoption study

11

2

?

2

?

2005-2007: staff member PhD scholar CLSU

Objectives CPWF Project Report

9. IMPACTS

9.1. Science

At the time of writing, four international journal papers have been published and seven are

submitted or in advanced stage of preparation (see chapter “Publications”). Three international

journal papers included parts of our CPWF results, and seventeen national journal papers, book or

proceeding chapters are published. Moreover, a large number of posters (>20) have been

presented, and oral presentations made at workshops, conferences, and other fora. We organized

an international Aerobic Rice workshop in Beijing, October 22-25, 2007, which attracted about 100

participants. All abstracts and papers are available on CDROM. The concept of aerobic rice is

getting accepted in scientific research as evidenced by participants with papers to the international

Aerobic Rice workshop in Beijing who were not part of our project.

9.2. Capacity building

Twenty-two graduate and undergraduate students were involved in the project, and most of them

have completed their theses (see chapter “Publications”). Fourteen students were enrolled at China

Agricultural University, three at Central Luzon State University (Philippines), two at the University

of the Philippines Los Banos (Philippines), two at Wageningen University (Netherlands), and one at

the Universita degli Studi di Firenze (Italy).

Aerobic rice was part of many training courses on water management in rice production systems

organized conjunctively by the Water Workgroup of the Irrigated Rice Research Consortium and

our CPWF project (we can not separate out the contributions as these were real joint activities).

The audiences of these courses were staff from institutes, universities, extension agencies, and

irrigation system administrators with a mandate for applied research, water management, or

extension. The trainings did not extent to our own project partners as they received “on the job

training”. These trainings were part of our outreach activities on aerobic rice. Table 9.2.1 lists the

most relevant course where aerobic rice was a major component, with number of participants. The

majority of the courses was organized in the Philippines, but there were also trainings in

Bangladesh (IGP), Vietnam (Mekong, Red River), and Myanmar (Irrawady). In total, 1589

professionals received training on aerobic rice during the lifetime of our CPWF project.

Table 9.2.1. Water management trainings with a significant component of aerobic rice, jointly

organized by the CPWF project and the IRRC, 2004-2007.

Course Title Place/Country

Date

offered

Partici-

pants

1

Integrated water management in

rice production -- technology

transfer for water savings

Los Banos,

Philippines

4-8

October

2004 34


Page | 109

2




Muñoz,

Philippines

8-12

November

2004 34

3




Tuguegarao,

Philippines

11-13

November

2004 38

4




Muñoz,

Philippines

8-10

December

2004 33

5



transfer for water savings Batac, Philippines

15-17

March

2005 40

6




Tarlac,

Philippines

15-17

March

2005 40

7

Training on Water Saving

Technologies

Pilar, Bohol,

Philippines

26-27 April

2005 35

8


Technologies

Barangay Patong,

Philippines

10-13 May

2005 35

9

Training on Integrated Field

Water Management Bagan, Myanmar

5-7

October

2005 40

10

Training on Integrated Field

Water Management Bohol, Philippines

19-21

December

2005 30?

11


Technologies Bohol, Philippines

Feb 2-17,

2006 500

12

IRRI Rice Production Training

Course

Los Banos,

Philippines 30-Mar 20

13

Training on water saving

technologies and implementation

of Integrated Crop Management

in Vietnam

Nan Dinh and

Habac, Vietnam

March 18-

26, 2006 30


Page | 110

14


Technologies

Manila,

Philippines

March 2-4,

2006

15

Training on component

technologies… Bohol, Philippines

July 5-6,

2006 24

16


Technologies Cebu, Philippines

18-23 July

2006 60

17


Technologies Cebu, Philippines

26-27 Sept

2006 40

18

Integrated field water

management

Negros,

Philippines

28-29 Sep

2006 40

19

Science and Technology updates

Rice production

Muñoz,

Philippines 26-Jan-07 50

20

Science and Technology updates

Rice production

Muñoz,

Philippines 08-Feb-07 50

21

Lecture: Water savings at NIA

Facilitators training

Quezon City,


22

Lecture: Water savings at NIA

Facilitators training

Quezon City,


23

Training workshop on water-

saving technologies for rice

production in rainfed areas

Bulacan,

Philippines 12-Apr-07 30

24

Water Management training

course Hanoi, Vietnam 03-May-07 50

25 Technology forum

Alaminos,

Pangsinan,

Philippines

may 10-

11, 2007 35

26 Water Management seminar

Los Banos,

Philippines 17-May-07 25

27


management

Maasin, Leyte,

Philippines

June 21-

22, 2007 40

28


management

Libmanan, Cam

Sur, Philippines

August 20-

21, 2007 36


Page | 111

29


management

Gazipur,

Bangladesh

August 26-

28, 2007 40

30 Water Management seminar IRRI, Los Banos 17-May-07 25

31


course

FCRI, Hanoi,

Vietnam

Dec 18

2007 50

32


course

NOMAFSI, Phuc

Ho, Vietnam

Dec 20

2007 15

Our national and local project partners organized many farmer trainings through farmer school

days and visits to demonstration sites. A typical farmer school day/demonstration day would

involve 50-100 farmers for the duration of one day. We did not keep track of all farmer school

days/demonstrations organized during the project as many were local initiatives that went

unrecorded. Table 9.2.2 provides a rough estimate of yearly outreach events and number of

farmers reached for each partner country. Over the 3.5 year span of our project, we estimate that

we reached 1875-3750 farmers with information about aerobic rice.

Table 9.2.2. Estimated yearly number of outreach events to farmers on aerobic rice, and estimated

number of farmers reached.

Country Area (province, county)

Participants

per site

Total

number

China

Fengtai, Mencheng, Kaifeng,

Funan, Fengyan 50-100 250-500

India Bulandshahar 50-100 50-100

Thailand

Scattered villages in NE

Thailand 50-100 50-100

Laos Three breeding trial sites 25-50 75-150

Philippines

Bulacan, Nueva Ezija, Tarlac,

Bataac 50-100 200-400

625-1250

9.3. Community impacts

We studied the socio-economics of aerobic rice cropping in a case study in China and the results

are reported in the paragraph “Farmers’ practice”. The major impact of the aerobic rice technology

is that it offers farmers the choice to grow rice when water is too scarce to grow lowland rice. It

allows farmers to diversify their cropping system which will augment resilience and increase overall


Page | 112

sustainability. Adoption will (among others) depend on relative market prices the different crops

available to farmers are expected to fetch in a particular year, the degree of water scarcity,

availability of quality seeds and extension support, and – last but not least – farmers’ preference

to grow their own rice for self sufficiency rather than depending on markets. Aerobic rice is not a

system that increases yields and profitability of rice per se, and therefore is not expected to

increase farm income (again depending on the price of relative crop products). However, it can

contribute to farm household food security and can have impacts on the price of rice if large-scale

adoption takes place (see paragraph “Global impacts”).

We commissioned an external impact/adoption study of aerobic rice in northern China and

results will be made available to the CPWF (and general public) early 2008.

9.4. Global impacts

We collaborated with the Impact Project of the CPWF to develop scenario analyses to explore the

impact potential of aerobic rice. Although this work is still in progress, we include summary results

to date as reported by Rubiano et al., (2007), with the scenario analyses done by IFPRI (Claudia

Ringler, Tingju Zhu). The scenario analyses were carried out using the IMPACT-WATER model to

explore the potential impacts of the adoption of aerobic rice over a 20-year period. The

International Model for Policy analysis of Agricultural Commodities and Trade (IMPACT) was

created in the 1990s to address a lack of a long-term vision and consensus about the actions that

are necessary to feed the world in the future, reduce poverty, and protect the natural resource

base. IMPACT covers 32 commodities (which account for virtually all of world food production and

consumption), including all cereals, soybeans, roots and tubers, meats, milk, eggs, oils, meals,

vegetables, fruits, sugar and sweeteners, and fish in a partial equilibrium framework. It is specified

as a set of 43 country and regional-level supply and demand equations, which are linked through

trade. IMPACT-WATER is an extension of the IMPACT model by a Water Simulation Model (WSM)

for projections of water supply and demand. Water supply and demand, and food production are

assessed at the river basin scale, and food production is summed to the national level, where food

demand and trade are modelled. Currently, IMPACT-WATER has aggregated spatial scales of to a

total of 126 river basins, 115 countries and regions, and 281 so called “food-producing units”

(FPUs).

The run used the extrapolation domain areas reported in paragraph “Extrapolation domain” that

corresponded to a 70% or greater chance of finding similar conditions to the pilot sites of our

project. Four scenarios were run:

1. Rainfed-only adoption, no climate change

2. Rainfed and irrigated adoption, no climate change

3. Rainfed-only adoption, climate change

4. Rainfed and irrigated adoption, no climate change


Page | 113

The following assumptions were made on adoption:

Rainfed-only adoption assumes:

a) The adoption of aerobic rice in rainfed areas starts from 2005 if the FPU includes pilot sites

or 2010 otherwise. Adoption continues for 20 years until the area grown to aerobic rice

equals the potential extrapolation domain in that FPU.

b) There is no adoption of aerobic rice in irrigated areas

c) Adoption of aerobic rice is assumed to increase rice yield by 50%, and the crop ET

requirement over the whole growing period are reduced to 450 mm for FPUs in temperate

climate zone, 900 mm in subtropical climate zone, and 1000 mm in tropical climate zone,

on average, varying from year to year according to climate conditions.

Rainfed and irrigated adoption assumes:

a) If the extrapolation domain (ED) area is less than the existing area of rainfed rice, only

rainfed rice areas will adopt.

b) If the (ED) area is greater than the rainfed rice area in the FPU, aerobic rice will replace all

the rainfed rice and part of the irrigated rice.

c) If the ED area is more than the current rice growing area in a FPU then aerobic rice will be

adopted over the whole are and there will be an expansion in the area grown under rainfed

rice.

The main results are:

Water savings. Analysis of potential water saving was conducted only for the adoption of aerobic

rice in irrigated areas, thus saving is only discussed for rainfed and irrigated adoption scenarios.

With climate change, adoption of aerobic rice could lead to a 50% saving in irrigated water in the

Yellow River, China, 28% saving in the Mekong, Thailand and about 22% average saving in four

FPUs in India. Without climate change, the only real saving in irrigated water adoption will be in

China. This is because under normal climatic conditions aerobic rice hardly affects ET, rather

water savings come from reduced percolation and seepage, factors that the IMPACT-WATER does

not model for. More water is saved under climate change because requirements for irrigated rice

increase with higher ET levels in the climate change scenario. Most of the water saving is in the

Indo-Gangetic basin.

Rice supply, demand, prices, and levels of malnutrition. Significant replacement of rainfed rice with

aerobic rice in the extrapolation domain areas will lead to changes in rice supply, demand, prices,

trade, and levels of malnutrition. Adoption of aerobic rice will have a greater positive impact under

climate change because climate change will increase water shortages, and adoption of aerobic rice

reduces the water requirement for rice production. The impact of aerobic rice on price production

will be concentrated in India and Thailand, which contain 44% and 33% of the potential aerobic

rice extrapolation domain areas respectively. Climate change will lead to increases in rice prices


Page | 114

over the next 50 years but adoption of aerobic rice will dampen the price increase with and without

climate change. Declining rice prices will make food more affordable for the poor. Without climate

change, the number of malnourished children declines from 147 million children in 2000 to 77

million children by 2050, mostly in South Asia and Sub-Saharan Africa. With climate change, the

decline is only to 81 million children by 2050. Adoption of aerobic rice through enhanced yield and

production will reduce malnutrition under both alternative scenarios.

Trade. India is the country with the largest potential expansion for tropical aerobic rice. According

to the IMPACT-WATER projections, India is a net rice importer by 2050 with net imports of 15

million metric tons without climate change, and 22 million metric tons with climate change. If rice

production in rainfed areas increases as a result of higher aerobic rice yields, then this will help

reduce rice imports. Additional production as a result of adoption of aerobic rice will be important

under climate change, because IMPACT-WATER predicts that there will be less volume of rice

available for trading in 2050 as rice production will be stressed in most developing-country

producers.

In summary, the analysis showed that the potential impacts of aerobic rice could be large and

beneficial, and will be even more important given temperature and water stress from increased

climate variability and climate change. The analysis also showed that most of the impacts would

occur in India and Thailand and would include increased yields, production, and area grown to rice,

reduced rice prices, and reduced levels of child malnutrition. The IMPACT-WATER model founds

that India will be a net importer of rice in 2050. Adoption of aerobic rice would reduce the amount

of imports necessary, which could be very important if climate change stresses rice-growing areas

in the tropics, as predicted, and reduces the amount of rice available for trade.

The results of the scenario analysis are sensitive to the evapotranspiration (ET) levels chosen for

aerobic rice. An earlier run used an ET of 450mm in all extrapolation areas and as a result

predicted much higher levels of water saving and yield increase. However, the main water savings

from aerobic rice come from reduced seepage and percolation losses, which is not yet factored into

the scenario analysis. Hence, the current predictions of water saving and yield increase on an

acreage basis are likely to be conservative.


Page | 115

10. PARTNERSHIP ACHIEVEMENTS

None of the reported results, outputs, outcome, or impacts would have been possible without the

partnerships established within the project and by the project with “new” partners (not listed as

partners in the original project proposal and signed contract). All achievements are a direct result

of the CPWF project partnership. Most of the reported results were obtained through activities

(experiments, modeling, surveys) by the project partners (Chapter “Results and outcomes); many

of the trainings were organized by or through project partners (paragraph “capacity building”); all

of the farmer schools, demonstration activities and other farmer-outreach events were organized

by the project partners (paragraph “capacity building”). Five universities contributed to graduate

and undergraduate training of project scholars. Important other partnership issues are reported in

chapter “Outcomes (partner change)”.

In the spirit of promoting partnerships of the CPWF, we strengthened existing links and

created new ones with other projects, programs, or consortia relevant to the development and

dissemination of aerobic rice (“network weaving”). Most importantly:

4. The water Workgroup of the Irrigated Rice Research Consortium (IRRC), that initiated the

first (modest) partnership on aerobic rice in 2001. Our CPWF project collaborated with a

number of the same partners and build-on and expanded their activities in China, India,

and the Philippines.

5. The Consortium for Unfavorable Rice Environments (CURE). We collaborated with this

consortium in the rainfed and more unfavorable environments of Laos and Thailand.

6. The ADB-funded project on “Developing and Disseminating Water-Saving Rice

Technologies in South Asia”. This project has a large component on aerobic rice and we

collaborated especially on the issue of soil health and sustainability. There are no common

project partners or sites (their sites are in India, Bangladesh, Pakistan, and Nepal).

We shared information (research protocols, ideas, results, outputs, data) as much as possible and

conducted many joint project planning and discussion meetings and workshops. In fact, there was

no CPWF project planning meeting or workshop that was not combined with one or more of the

other project/consortia meetings. The advantages of this collaboration to the development of the

aerobic rice technology can hardly be quantified (or overstated!), but all participants considered

them of tremendous benefit. The only drawback may be the difficulty in attributing certain

activities, results, outcomes, or impacts specifically to any of the partnership members (such as

the case of trainings reported in paragraph “Capacity building”). For the overall goal of poverty

alleviation and food security, such considerations, however, should not matter.


Page | 116

11. RECOMMENDATIONS

The aerobic rice technology can be considered sufficiently mature for dissemination in the Yellow

River Basin (YRB), but not so for most of tropical Asia where more R&D is needed to create

sustainable and high-yielding aerobic rice systems. Therefore, our recommendations are made for

two regions: the YRB, and “Tropical Asia”. The IGP in central India seems to take an intermediate

position.

11.1. Policy makers

We recommend that “aerobic rice” be recognized as a special crop type besides the existing

categories of “lowland rice and “upland rice”. This will facilitate the collection of statistics on

adoption and spread of aerobic rice, and will encourage formal extension agencies to address

aerobic rice and meet farmers’ demands for extension support. In the YRB, local and national

governments should strengthen the extension system at all levels. To date, most knowledge on

aerobic rice is located within universities and institutes and extension agencies have little

knowledge of aerobic rice. In most of tropical Asia, preference should be given to setting up

dedicated aerobic rice breeding programs and strengthening the R&D capacity to develop

sustainable production systems.

11.2. Extension

In the YRB, the formal extension agencies should be equipped with knowledge on aerobic rice.

However, the “informal” extension sector is a powerful agent of change and should be included in

the supply of information. We usually found local “technicians”, i.e. agricultural shop keepers and

informal farmer leaders with some agricultural training, at the basis of adoption of aerobic rice

whenever we interviewed farmers.

In most of tropical Asia, care should be taken with the extension of aerobic rice. Although

yields of current aerobic varieties can reach up to the same level as in the YRB, they are not as

well adapted to aerobic soil conditions and the soil needs to be kept much wetter (using more

water) than in the YRB. Moreover, the experiences with “yield collapse” in the Philippines suggest

that more R&D effort is needed to develop high-yielding and sustainable systems. In the

Philippines, we encountered yield collapse in some farmers’ fields (e.g., Tarlac, Nueva Ezija), but

we also had experiences of high yields in others with no sign of yield collapse at all (e.g., Bulacan).

Where water is scarce, farmers are usually very keen to try aerobic rice and careful guidance by

extension agents is warranted. In central India (IGP), lowland rice is often grown on permeable

soils with intermittent flooding (because of water scarcity). Here, aerobic rice yields were found to

be at par with such lowland rice systems, but using less water. Aerobic rice can be disseminated in

such areas, using selected varieties as reported here, as no cases of yield collapse have been

reported. In Thailand and Laos (Mekong), more breeding and variety selection needs to take place

before aerobic rice can be promoted on a large-scale. Also, improved management practices need

first be developed.


Page | 117

11.3. Research

At the international Aerobic Rice workshop (Beijing, 22-25 October 2007), three independent

experts presented a detailed list of research recommendations: dr E. Humphreys (CPWF-Theme

Leader Crop Water Productivity), dr G. Singleton (IRRC coordinator), and dr A. Dobermann

(program leader irrigated environment, IRRI). These recommendations are completely and without

editing reproduced in Appendix C. Based on these recommendations and our own experiences, we

propose the following major recommendations:

General (both YRB and the Tropics)

1. A better understanding is needed of the factors that influence farmers in adopting/adapting

an aerobic rice cropping system. Impact and adoption studies are needed, especially in

areas where aerobic rice is currently being adopted such as northern China. Such studies

should include both biophysical factors (eg water availability, soil type) and socio-economic

factors.

2. Long-term trials such as at IRRI need to be established in a number of target locations to

study long-term sustainability and develop sustainable practices. Does yield decline with

continuous cropping? Are there any soil-health issues? What are suitable crop rotations?

What will be the pest and disease problems in aerobic systems (above ground and below

ground)?

3. Make an inventory of “soil health” threats across the potential target domains.

YRB

1. Improve the yield potential. Highest recorded yields in our field experiments were 6 t ha-1,

whereas simulation modeling suggests that yields could go up to 8 t ha-1. We need to

explore breeding options to increase the panicle (sink) size since the source size of current

aerobic rice varieties seems strong enough. Also, we need to further investigate whether

yield of current varieties can be increased by improved management, especially that of

micro-nutrients such as manganese. There seems to be a lot of scope to improve

management, especially related to plant nutrition.

2. Explore the potential impact of large-scale adoption of aerobic rice in regional hydrology

and water availability. How much water is saved on regional basis (up to now, we focused

on field scale only)? We need actual measurements of evapotranspiration (ET) rates of

aerobic rice (up to know, we mostly estimated ET by simulation modeling and by water

balance approaches).

Tropics

1. Varieties need to be developed with higher tolerance to aerobic soil, aiming at yield

potentials of 5-6 t ha-1 with soil water tensions going up to 100 kPa (like with the Han Dao

aerobic rice varieties in the YRB). Besides improving the tolerance to aerobic soil


Page | 118

conditions, tolerance to “soil sickness”, such as nematodes, fungi, or nutrient imbalances

may be important in increasing yield potential as well.

2. The mechanisms of yield decline under continuous cropping and of immediate yield

collapse need to be understood and remedial (management) practices developed. It is

proposed to develop specifically sustainable crop rotation systems.


Page | 119

12. KEY PUBLICATIONS

Published Journal papers

1. Bouman, B.A.M., Feng Liping, Tuong, T.P., Lu Guoan, Wang Huaqi, Feng Yuehua, 2007.

Exploring options to grow rice under water-short conditions in northern China using a

modelling approach. II: Quantifying yield, water balance components, and water

productivity. Agricultural Water Management 88, 23-33.

2. Feng Liping, Bouman, B.A.M., Tuong, T.P., Cabangon, R.J., Li Yalong , Lu Guoan, Feng

Yuehua, 2007. Exploring options to grow rice under water-short conditions in northern

China using a modelling approach. I: Field experiments and model evaluation. Agricultural

Water Management 88, 1-13.

3. Lixiao Nie, Shaobing Peng, Bas A. M. Bouman, Jianliang Huang, Kehui Cui, Romeo M.

Visperas, and Hong-Kyu Park, 2007. Solophos fertilizer improved rice plant growth in

aerobic soil. Journal of Integrated Field Science 4: 11-16.

4. Lixiao Nie, Shaobing Peng, Bas A.M. Bouman, Jianliang Huang, Kehui Cui, Romeo M.

Visperas, Hong-Kyu Park, 2007. Effect of oven heating of soil on vegetative plant growth of

aerobic rice. Plant and Soil 300: 185-195.

Journal papers submitted or in preparation

1. Lampayan, R.M., B.A.M. Bouman, J.E. Faronilo, J.B. Soriano, L. B. Silverio, B. V.

Villanueva, J.L. de Dios, A.J. Espiritu, T.M. Norte, K. Thant, (in prep 2008), Yield of aerobic

rice in rainfed lowlands of the Philippines as affected by N fertilization and row spacing.

2. Limeng Zhang, Shan Lin, Hongbin Tao, Changying Xue, Xiaoguang Yang, Dule Zhao,

B.A.M. Bouman, Klaus Dittert (in prep 2008), Response of yield determinants and dry

matter translocation of aerobic rice to N application on two soil types

3. Lixiao Nie, Shaobing Peng, Bas A.M. Bouman, Jianliang Huang, Kehui Cui, Romeo M.

Visperas, and Jing Xiang, (in preparation 1), Alleviating soil sickness caused by aerobic

monocropping: Responses of aerobic rice to nutrient supply, to be submitted to Field Crops

Research.

4. Lixiao Nie, Shaobing Peng, Bas A.M. Bouman, Jianliang Huang , Kehui Cui, Romeo M.

Visperas, Jing Xiang, (in preparation 2), Alleviation of Soil Sickness Caused by aerobic

monocropping: Responses of rice plant to various nitrogen sources, to be submitted to

Plant and Soil.

5. Xiao Qin Dai, Hong Yan Zhang, J. H. J. Spiertz, Jun Yu, Guang Hui Xie, B. A. M Bouman, (in

prep 2008), Productivity and Resource Use of Aerobic Rice – Winter Wheat Cropping

Systems in the Huai River Basin.


Page | 120

6. Xue Changying, Yang Xiaoguang, Bouman B.A.M., Deng Wei, Zhang Qiuping,Yan Weixiong,

Zhang Tianyi, Rouzi Aji, Wang Huaqi,Wang Pu, (submitted), Effects of irrigation and

nitrogen on the performance of aerobic rice in northern China. Journal of Integrative Plant

Biology.

7. Xue Changying, Yang Xiaoguang, B.A.M. Bouman, Deng Wei, Zhang Qiuping, Yan

Weixiong, Zhang Tianyi, Rouzi Aji, Wang Huaqi, (submitted), Optimizing yield, water

requirements, and water productivity of aerobic rice for the North China Plain, Irrigation

Science.

Other scientific publications

National Journal publications, proceedings, book chapters. Some papers report only on CPWF

project results, whereas in others, CPWF project results are part of the paper. We included

international journal papers which include part of our CPWF project results as well.

1. Guanghui Xie, Yu Jun, Yan Jing, Wang Huaqi, Zhu Xiurong, 2007. Direct seeding of aerobic

rice in China. In: Rice is life: scientific perspectives for the 21st century. IRRI, Los Banos,

pp 186-188.

2. Bouman, B.A.M., Yang Xiaoguang, Wang Huaqi, Wang Zhimin, Zhao Junfang, Chen Bin,

2006. Performance of aerobic rice varieties under irrigated conditions in North China. Field

Crops Research 97, 53-65.

3. Bouman, B.A.M., Humphreys, E., Tuong, T.P., Barker, R., 2006. Rice and water. Advances

in Agronomy 92, 187 - 237.

4. Bouman B.A.M, Barker R., Humphreys E., Tuong T.P., Atlin G.N., Bennett J., Dawe D.,

Dittert K., Dobermann, A., Facon T., Fujimoto N., Gupta R. K., Haefele S.M., Hosen Y.,

Ismail A.M., Johnson D., Johnson S., Khan S., Lin Shan, Masih I., Matsuno Y., Pandey S.,

Peng S., Thiyagarajan T.M., Wassman R., 2007. Rice: feeding the billions. In: Water for

Food, Water for Life: A Comprehensive Assessment of Water Management in

Agriculture. London: Earthscan and Colombo: International Water Management Institute.

Pp 515-549.

5. Lampayan, R.M., B.A.M. Bouman, 2005. Management strategies for saving water and

increase its productivity in lowland rice-based ecosystems. In: proceedings of the First

Asia-Europe Workshop on Sustainable Resource Management and Policy Options for Rice

Ecosystems (SUMAPOL), 11-14 May 2005, Hangzhou, Zhejiang Province, P.R. China. On

CDROM, Altera, Wageningen, Netherlands.

6. Liping Feng, Bouman, B.A.M., Tuong, T.P., Yalong Li, Guoan Lu, Cabangon, R.J., Yuehua

Feng, 2006. Effects of groundwater depth and water-saving irrigation on rice yield and

water balance in the Liuyuankou Irrigation System, Henan, China. In: Agricultural water

management in China, Willet, I.R., Gao Zhanyi (eds.), ACIAR Proceedings 123, Canberra,

Australia. Pp 52-66.


Page | 121

7. Peng, S., Bouman, B.A.M., 2007. Prospects for genetic improvement to increase lowland

rice yields with less water and nitrogen. In: Spiertz, J.H.J., Struik, P.C., Van Laar, H.H.

(Eds.), Scale and complexity in plant systems research – Gene-plant-crop relations.

Springer, Dordrecht, Netherlands. Pp 251-266.

8. Peng Shaobing, Bouman, B.A.M., Visperas, R.M., Castañeda, A., Nie Lixiao, Park Hong-

Kyu, 2006. Comparison between aerobic and flooded rice in the tropics: agronomic

performance in an eight-season experiment. Field Crops Research 96, 252-259.

9. Soriano, J.B., Valdez, J.A., Silverio, L.B., Pastrana, M.I., Cabiles, D.M.S., Mendoza, J.P.,

Bouman, B.A.M., Lamapayan, R.M., Villanueva, B.V., Peralta, W., 2006. Aerobic rice

technology in rainfed areas of Bulacan: A R&D project. CLARRDEC Science and technology

Journal 1: 29-46.

10. Xue Chang-Ying, Yang Xiao-Guang, Bouman, B.A.M., Feng Li-Ping, Gon van Laar, Wang

Hua-Qi, Wang Pu, Wang Zhi-Min, 2005. Preliminary Approach on Adaptability of

ORYZA2000 model for aerobic rice in Beijing region. Acta Agronomica Sinica, 31(12),

1567-1571 (Chinese with English abstract).

11. Xue Changying ,Yang Xiaoguang, Deng Wei, Zhang Tianyi, Yan Weixiong, Zhang Qiuping,

Rouzi Aiji, Zhao Junfang, Yang Jie, Bouman. B.A.M, 2007. Yield Potential and Water

Requirement of Aerobic Rice in Beijing Analyzed by ORYZA2000 Model, Acta Agronomica

Sinica, Vol 33(4): 625-631.

12. Xue Changying ,Yang Xiaoguang, Deng Wei, Bouman. B.A.M, Zhang Qiuping, Yan

Weixiong, Zhang Tianyi, Rouzi Aiji, Wang Huaqi, 2007. Study on Yield Potential and Water

Requirement of Aerobic Rice in Beijing Area Based on ORYZA2000 Model. In: Hu

Yuegao,Huang Guohe, Li Zhaohu (Eds), Principles and Practices of Desertification Control,

Proceedings of the international specialty conference on science and technology for

desertification control. China Meteorological Press, Beijing, China. Pp: 435-451.

13. Yang Jie, Yang Xiaoguang, Wang Huaqi, Wang Pu, B.A.M. Bouman, 2005. Soil water

characteristics of farmland of aerobic rice. Chinese Journal of Eco-Agriculture, 11(3), 82-

86.

14. Yang Xiaoguang, B.A.M. Bouman, Zhang Qiuping, Xue Changying, Zhang Tianyi, Xu

Jianyong, Wang Pu, Wang Huaqi, 2006. Crop coefficient of aerobic rice in North China.

Transactions of the Chinese Society of Agricultural Engineering 22,(2), 37-41 (Chinese with

English abstract).

15. Zhang Qiuping, Yang Xiaoguang, Yang Jie, Wang Huaqi, Wang Pu, Wang Zhimin,

B.A.M.Bouman, 2005. The Studies of Photosynthesis Characteristics and Water Use

Efficiency on Aerobic Rice under Different Irrigation Treatments. Transactions of the

Chinese of Agricultural Research in the Arid Areas.Vol.23, No.4: 67-72.

16. Zhang Qiuping, Yang Xiaoguang, Xue Changying, Yan Weixiong, Zhang Tianyi, Bouman.

B.A.M, Wang Huaqi, 2007. Analysis of Coupling Degree between Crop Water Requirement


Page | 122

of Aerobic Rice and Rainfall in Beijing Aereas, Transactions of the Chinese Society of

Agricultural Engineering, Vol. 23 (10):51-56.

Student dissertations

1. Xue Changying, Evaluation of ORYZA2000 Model or Aerobic Rice and Studies on

Optimization of Irrigation, MsC thesis, China Agricultural University. June 2005.

2. Zhang Qiuping, Studies on the Irrigation Model of Aerobic Rice under Beijing Climate

Background, MsC thesis, China Agricultural University. June 2005.

3. Yan Weixiong, Effects of Different Water and Nitrogen Regimes on Yield and Quality

Formation of Aerobic Rice, MsC thesis, China Agricultural University. June 2006.

4. Rouzi Aji, Photosynthetic Characteristics and Water Use Efficiency of Aerobic Rice under

Different Irrigation Treatments, MsC thesis, China Agricultural University. June 2006.

5. Zhang Tianyi, Analysis for Soil Water Balance of Aerobic rice in Beijing based on

ORYZA2000 model, MsC thesis, China Agricultural University. June 2006.

6. Xue Changying, Water use and water productivity of aerobic rice under different water x

nitrogen regimes. PhD thesis, China Agricultural University, 2005-2008.

7. Zhang Limeng, Yield Formation and Nitrogen Uptake of Aerobic Rice Response to N

Application and Irrigation in North China, PhD thesis, China Agricultural University, 2005-

2008.

8. Liu Po, Nitrogen Uptake of Aerobic Rice HD297 in Response to Nitrogen Fertilization and

Irrigation, MsC thesis, China Agricultural University, 2007.

9. Ruan Kanglei, Manganese Uptake of Aerobic Rice HD297 in Response to Mn Fertilization

and Irrigation, MsC thesis, China Agricultural University, 2007.

10. Zhang Hongyan, N,P,K, Uptakes and Soil Indigenous Supplies to Aerobic Rice and Rotated

Winter Wheat at Mengcheng, MsC thesis, China Agricultural University, 2007.

11. Tao Guangcan, A Comparison of Yield Formation and Nutrient Uptakes of Aerobic Rice in

Different Climatic Conditions, PhD thesis, China Agricultural University 2007.

12. Yu Jun, Tiller Contribution to Yield of Aerobic Rice, MsC-PhD thesis, China Agricultural

University, 2008

13. Yan Jing, Nitrogen Use and Farming Systems based on Aerobic Rice in Mengcheng, PhD

thesis, China Agricultural University, 2006-2010.

14. “Rice Farming with Shoes on’: Strategy, Decision-making and Performance in Farmer

Adoption of Aerobic Rice in Pala-pala, Bulacan, Philipines”, Rica Flor, MsC thesis University

of the Philippines, Los Banos, 2007.

15. Water and nutrient management for aerobic rice (Oryza sativa L.) production system.

Junel B. Soriano. PhD Thesis, Central Luzon State University, Munoz, 2008 (submitted for

pre-defense, January 2008).


Page | 123

16. The effect of different water regime and nitrogen application rate on aerobic rice grown

under CLSU condition. Von Eliel Bauzo Camaso Eleazar Valdez Ranese Jr. BsC thesis,

Central Luzon State University, Munoz, 2006.

17. Performance of different rice varieties as influenced by water regimes under aerobic

conditions. Khin Mar Htay, PhD Thesis, Central Luzon State University, Munoz, 2007.

18. Bending resistance and yield of aerobic rice (Oryza sativa L.) as affected by nitrogen

application and row spacing. Khin Myo Thant. MsC thesis University of the Philippines, Los

Banos, 2005.

19. A pot experiment on ‘soil sickness’ in aerobic rice: A biotic problem? Hanneke Vermeulen.

MsC thesis, Wageningen University, Wageningen, 2007.

20. Evaluation of the Aerobic Rice technology: three years of experiments in the Philippines.

Hanneke Vermeulen. Internship thesis, Wageningen University, Wageningen, 2007.

21. Evaluation of the Aerobic Rice technology: two years of experiments in the Philippines.

Matthijs Bouwknegt. MsC thesis, Wageningen University, Wageningen, 2005.

22. The aerobic rice project: evaluation of the water use efficiency in rice-growing. Grassi

Chiara, MsC Thesis, Universita degli Studi di Firenze, Florence, Italy, 2006.

Popular articles (English)

1. A “Grain of Truth” on aerobic rice in the journal Rice Today (Published by IRRI), June

2007.

2. An article on aerobic rice in Ripple, June 2007 (IRRC newsletter)

3. A feature article in Rice Today (special issue October 2007): “Aerobic rice farming with less

water —fighting the irrigation crisis”, by Adam Barclay.

Training materials

Bouman, B.A.M., R.M., Lampayan, T.P. Tuong, 2007. Water management in Rice: coping with

water scarcity. Los Baños, (Philippines): International Rice Research Institute, 54 pp. The chapter

on aerobic rice was contributed by the CPWF project

Powerpoint presentation as part of training series on water management: aerobic rice module. A

number of Chinese-language presentations and information leaflets. Various flipcharts, posters,

leaflets produced by local project partners as used in farmer field schools and other training

events.


Page | 124

BIBLIOGRAPHY

Anderson BH, Magdoff FR (2005) Autoclaving soil samples affects algal-available phosphorus. J

Environ Qual 34:1958-1963.

Bouman, B.A.M. and T.P. Tuong, 2001. Field water management to save water and increase its

productivity in irrigated rice. Agricultural Water Management, 49/1: 11-30.

Bouman BAM, Kropff MJ, Tuong TP, Wopereis MCS, Ten Berge HFM, Van Laar HH ( 2001)

ORYZA2000: modeling lowland rice. International Rice Research Institute, Los Baños, Philippines,

and Wageningen University and Research Centre, Wageningen, Netherlands, 1.

Bouman, B.A.M., Peng, S., Castañeda, A., Visperas, R.M., 2005. Yield and water use of irrigated

tropical aerobic rice systems. Agric. Water Manage. 74, 87-105.

Bouman, B.A.M., Humphreys, E., Tuong, T.P., Barker, R., 2006. Rice and water. Advances in

Agronomy 92, 187 - 237.

Bouman, B.A.M., Yang Xiaoguang, Wang Huaqi, Wang Zhimin, Zhao Junfang, Chen Bin, 2006.

Performance of aerobic rice varieties under irrigated conditions in North China. Field Crops

Research 97, 53-65.

Bouman, B.A.M., R.M., Lampayan, T.P. Tuong, 2007. Water management in Rice: coping with

water scarcity. Los Baños, (Philippines): International Rice Research Institute, 54 pp.

Cruz RT, and O’Toole JC (1984) Dryland rice response to an irrigation gradient at flowering stage.

Agron J 76: 178-183.

Doorenbos, J., Pruit, W.O., 1984. Guidelines for predicting crop water requirements. FAO Irrigation

and Drainage paper 24, Food and Agriculture Organization, Rome (Italy), p. 144.

Ekanayake IJ, De Datta SK, Steponkus PL (1989) Spikelet sterility and flowering response of rice

to water stress at anthesis. Annals of Botany 63: 257-264.

George T., Magbanua R., Garrity D.P., Tubaña B.S., and Quiton J. 2002. Rapid yield loss of rice

cropped successively in aerobic soil. Agron. J. 94: 981-989.

Hafeez MM, Bouman BAM, Van de Giessen N, Vlek P (2007) Scale effects on water use and water

productivity in a rice-based irrigation system (UPRIIS) in the Philippines, Agric Water Manage 92:

81-89.

Kirkegaard JA, Munns R, James RA, Gardner PA, Angus JF (1995) Reduced growth and yield of

wheat with conservation cropping. 2. Soil biological factors limit growth under direct drilling.

Austra J Agric Res 46(1):75-88.

Lin, S., Dittert, K., Sattelmacher, B., 2002. The Ground Cover Rice Production System (GCRPS)—a

successful new approach to save water and increase nitrogen fertilizer efficiency? In: Bouman,

B.A.M., Hengsdijk, H., Hardy, B., Bindraban, P.S., Tuong, T.P., Ladha, J.K. (Eds.), Water-wise rice

production. Proc International Workshop on Water-wise Rice Production, 8-11 April, Los Baños,

Philippines. International Rice Research Institute, Los Baños, Philippines, pp. 187-196.


Page | 125

Liu, X., Ju, X., Zhang, F., Pan, J., Christie, P. 2003. Nitrogen dynamics and budgets in a winter

wheat-maize cropping system in the North China Plain. Field Crops Res. 83, 111-124.

Liu, X., Ju, X., Zhang, Y., He, C., Kopsch, J., Zhang F. 2006. Nitrogen deposition in

agroecosystems in the Beijing area. Agriculture, Ecosystems and Environment. 113, 370-377.

Loeve R, Dong B, Molden D, Li YH, Chen CD, Wang JZ (2004) Issues of scale in water productivity

in the Zhanghe irrigation system: implications for irrigation in the basin context. Paddy and Water

Environment 2: 227-236.

Ma, W.Q. 1999. Current status and evaluation of crop fertilization in Shandong province. Ph.D.

Thesis. China Agricultural University, Beijing, China (Chinese with English abstract).

Molden D, Murray-Rust H, Sakthivadivel R, Makin I (2003) A water productivity framework for

understanding and action In: Kijne, J.W., Barker, R., Molden, D. (Eds.), Water productivity in

agriculture: limits and opportunities for improvement, CABI Publishing Wallingford, UK, pp 1-18.

Nishio, M., Kusano, S., 1975. Effect of root residues on the growth of upland rice. Soil Sci. Plant

Nutr. 21, 391-395.

Nishizawa T., Ohshima, Y., Kurihara H. 1971. Survey of the nematode population in the

experimental fields of successive or rotative plantation. Proc. Kanto-Tosan Plant Protection Soc.18:

121-122.

Peng, S., Bouman, B.A.M., Visperas, R.M., Castañeda, A., Nie, L., Park, H.K., 2006. Comparison

between aerobic and flooded rice in the tropics: agronomic performance in an eight-season

experiment. Field Crops Res. 96, 252-259.

Pinheiro B.D., de Castro E.D.M., Guimaraes C.M. 2006. Sustainability and profitability of aerobic

rice production in Brazil. Field Crops Research, 97 (1): 34-42.

Jorge Rubiano, Maya Rajasekharan, Claudia Ringler, Tingju Zhu, Boru Douthwaite, Simon Cook,

Bas Bouman, 2007. Impact Potential of the “Temperate and Tropical Aerobic Rice (STAR) in Asia”

Project (PN16), paper presented at the international Aerobic Rice Workshop, Beijing, October

2007.

Sasaki Y, Hosen Y, Peng S, Nie L, Rodriguez R, Agbisit R, Fernandez L, Bouman BAM, Kobayashi N

(2006) Possibility of abiotic factors on the gradual yield decline under continuous aerobic rice

cropping system. Paper presented at the Annual Meeting of Japanese Society of Soil Science and

Plant Nutrition, 5-7 September 2006, Akita, Japan, Vol. 52, pp 116.

Sun HY, Liu CM, Zhang XY, Shen YJ, Zhang YQ (2006) Effects of irrigation on water balance, yield

and WUE of winter wheat in the North China Plain. Agric Water Manage 85: 211-218.

Tao, H., T., Brueck, H., Dittert, K., Kreye, C., Lin, S., Sattelmacher, B. 2006. Growth and yield

formation of rice (Oryza sativa L.) in the water saving ground cover rice production system

(GCRPS). Field Crops Res. 95, 1-12.


Page | 126

Tuong, T.P., Bouman, B.A.M., Mortimer, M., 2005. More rice, less water – integrated approaches

for increasing water productivity in irrigated rice-based systems in Asia. Plant Production Sciences,

8(3): 229-239.

Van Ittersum, M.K., Rabbinge, R., 1997. Concepts in production ecology for analysis and

quantification of agricultural input-output combinations. Field Crops Research 52, 197-208.

Ventura W. and Watanabe I. 1978. Growth inhibition due to continuous cropping of dryland rice

and other crops. Soil Sci. Plant Nutr. 24: 375-389.

Ventura W., Watanabe I., Komada H., Nishio M., dela Cruz A., and Castillo M.B. 1984. Soil

sickness caused by continuous cropping of upland rice, mungbean, and other crops. IRRI Res

Paper Series 99, p 13.

Wang HX, Zhang L, Dawes WR, Liu CM (2001) Improving water use efficiency of irrigated crops in

the North China Plain-measurements and modeling. Agric Water Manage 48: 151-167.

Wang Huaqi, Bouman BAM, Dule Z, Wang C and Moya PF. (2002) Aerobic Rice in Northern China -

Opportunities and Challenges. In: Bouman BAM, Hengsdijk H, Hardy B, Bindraban PS, Tuong TP

and Ladha JK (Eds). Water-wise Rice Production. Proceedings of the International Workshop on

water-wise Rice Production, 8-11 April, 2002, Los Baños, Philippines, International Rice Research

Institute, pp. 143-154.

Wopereis MCS, Bouman BAM, Tuong TP, Berge HFM, Kropff MJ (1996) ORYZA_W: Rice growth

model for irrigated and rainfed environments. In: SARP Research Proceedings, IRRI/AB-DLO,

Wageningen, the Netherlands, pp15-54.

Yang Xiaoguang, Bouman, B.A.M., Wang Huaqi, Wang Zhimin, Zhao Junfang, Chen Bin, 2005.

Performance of temperate aerobic rice under different water regimes in North China. Agricultural

Water Management 74, 107-122.

Zhao, J.R., Guo, Q., Guo, J.R., Wei, D.M., Wang, C.W., Liu, Y., Lin, K. 1997. Investigation and

analysis on current status of chemical fertilizer inputs and crop yields in agricultural field of Beijing

Suburb. Beijing Agric. Sci. 15, 36–38 (Chinese).

Zi T (Ed.) (1999) Chinese Soil Taxonomy. Sciences Press, Beijing, pp 864.


Page | 127

PROJECT PARTICIPANTS

Partners with funding from CPWF

Name of Institution: International Rice Research Institute (IRRI)

Postal Address: DAPO Box 7777, Metro Manila, Philippines

Email: [email protected]

Type of Institution: CGIAR

Project members: B.A.M. Bouman (project leader), G. Atlin, Dule Zhao, Shaobing Peng, C. Kreye,

R. Lampayan, P. Moya, R. Bayot, R. Flor, L. Llorca, E. Quicho

Name of Institution: China Agricultural University (CAU)

Postal Address: No. 2 Yuangmingyuan Xilu Haidian District, Beijing 100094, P.R.

China,


Type of Institution: NARES

Project members: Wang Huaqi (PI), Yang Xiaoguang, Lin Shan, Xie Guanghui, Liping Feng, Tao

Hongbin, Zhang Limeng, Xue Changying, Yan Ying, Yu Jun, Xiao Qindai

Name of Institution: Indian Agricultural Research Institute-Water

Technology Centre

Postal Address: New Delhi 110 012, India



Project members: Anil Kumar Singh (PI), dr. Viswanathan Chinnusamy, S.K. Dubey

Name of Institution: National Irrigation Administration Tarlac

Postal Address: NIA Compound, Matatalaib, Tarlac 2300, Philippines



Project members: V. Vicmudo (PI), A. Lactauan, T. Norte

Name of Institution: Philippine Rice Research Institute (PhilRice)

Postal Address: Maligaya, 3119 Muñoz, Nueva Ecija, Philippines


Page | 128



Project members: J. de Dios, A. Espiritu

Name of Institution: Ubon Ratchathani Rice Research Center

Postal Address: P.O. Box 65, Ubon Ratchathani, 3400, Thailand



Project members: dr. B. Jongdee, dr. P. Konghakote (Khon Kean Rice Research Center)

Name of Institution: National Agriculture and Forestry Research Institute

Postal Address: Vientiane, Laos



Project members: dr Kouang Douangsilla, Mr. Sipaseuth

Name of Institution: Institute of Plant Nutrition and Soil Science

Postal Address: Christian Albrecht University Kiel (CAU-Kiel), Olshausenstr. 40,

24118 Kiel, Germany


Type of Institution: ARI

Project members: dr. K. Dittert

Contributing partners without funding from CPWF

Name of Institution: Central Luzon State University

Postal Address: Science City of Munoz, Philippines



Project members: dr A. Espino

Name of Institution: Bulacan Agricultural State College

Postal Address: San Ildefonso, Bulacan 3010, Philippines



Page | 129


Project members: dr J. Valdez, Mr. J. Soriano

Name of Institution: National Soil and Water Resources Research and Development

Center - Bureau of Soil and Water Management

Postal Address: San Ildefonso, Bulacan, Philippines

Email:


Project members: dr B.V. Villanueva


Page | 130

APPENDICES

Appendix A: Abstracts of all key publications

Bouman, B.A.M., Feng Liping, Tuong, T.P., Lu Guoan, Wang Huaqi, Feng Yuehua, 2007. Exploring

options to grow rice under water-short conditions in northern China using a modelling approach. II:

Quantifying yield, water balance components, and water productivity. Agricultural Water Management

88, 23-33.

Abstract

Because of increasing competition for water, water-saving technologies such as alternate wetting

and drying and aerobic rice are being developed to reduce water use while maintaining a high yield

of rice. The components of the water balance of these systems need to be disentangled to

extrapolate water savings at the field scale to the irrigation system scale. In this study, simulation

modelling was used to quantify yield, water productivity, and water balance components of

alternate wetting and drying and aerobic rice in the conjunctive surface-groundwater Liuyuankou

Irrigation System, Henan. The study on aerobic rice was supported by on-farm testing. In the

lowland rice area, where groundwater tables are within the root zone of the crop, irrigation water

savings of 200-900 mm can be realized by adopting alternate wetting and drying or rainfed

cultivation, while maintaining yields at 6400-9200 kg ha-1. Most of the water savings are caused

by reduced percolation rates, which will reduce groundwater recharge and may lead to decreased

opportunities for groundwater irrigation. Evaporation losses can be reduced by a maximum of 60-

100 mm by adopting rainfed cultivation. In the transition zone between lowland rice and upland

crops, groundwater tables vary from 10 cm to more than 200 cm depth, and aerobic rice yields of

3800-5600 kg ha-1 are feasible with as little as 2-3 supplementary irrigations (totaling 150-225

mm of water). Depending on groundwater depth and amount of rainfall, either groundwater

recharge or net extraction of water from the soil or the groundwater takes place.

Feng Liping, Bouman, B.A.M., Tuong, T.P., Cabangon, R.J., Li Yalong , Lu Guoan, Feng

Yuehua, 2007. Exploring options to grow rice under water-short conditions in northern

China using a modelling approach. I: Field experiments and model evaluation.

Agricultural Water Management 88, 1-13.

Abstract

China’s grain basket in the North China Plain is threatened by increasing water scarcity and there

is an urgent need to develop water-saving irrigation strategies. Water savings in rice can be

realized by alternate wetting and drying (AWD) under lowland conditions, or by aerobic rice in

which the crop is grown under nonflooded conditions with supplemental irrigation. Field


Page | 131

experimentation and simulation modelling are a powerful combination to understand complex

crop-water interactions and to extrapolate site-specific empirical results to other environments and

conditions. In this paper, we present results from four years of field experiments on AWD and

aerobic rice in 2001-2004 near Kaifeng, Henan Province, China. The experimental data were used

to parameterize and evaluate the rice growth model ORYZA2000. A subsequent paper reports on

the extrapolation of the experimental results using ORYZA2000 and on farmer-participatory testing

of aerobic rice. In the lowland area of the study site, rice yields under flooded conditions were

around 8000 kg ha-1 with 900 mm total (rain, irrigation) water input. Irrigation water savings were

40-70% without any yield loss by applying AWD. In the upland area of the study site, aerobic rice

yielded 2400-3600 kg ha-1, using 750-1100 mm total water input. ORYZA2000 satisfactorily

reproduced the dynamics in measured crop variables (biomass, leaf area, N uptake) and soil water

variables (ponded water depth, soil water tension). The root mean square error of predicted yield

was 11% for lowland rice and 19% for aerobic rice, which was only one and a half times the error

in the measured values. We concluded that ORYZA2000 is sufficiently accurate to extrapolate our

results on AWD and aerobic rice to different management and environmental conditions in our

study area.

R.M. Lampayan, B.A.M. Bouman, J.E. Faronilo, J.B. Soriano, L. B. Silverio, B. V.

Villanueva, J.L. de Dios, A.J. Espiritu, T.M. Norte, K. Thant, (in prep), Yield of aerobic rice

in rainfed lowlands of the Philippines as affected by N fertilization and row spacing.

Abstract.

This study evaluated the effects of different N amount and timing of N applications and row

spacings on yields of aerobic rice under rainfed conditions in Central Luzon, Philippines. Two field

experiments were conducted: (a) nitrogen amount by row spacing (NA x RS) experiment in San

Ildefonso, Bulacan and in Dapdap, Tarlac during 2004 and 2005 wet season, and (b) N-split

application by row spacing (NS x RS) in PhilRice during 2004 wet season and 2005 dry season. All

experiments were laid out in a split-plot design with four replications using Apo cultivar, and

nitrogen as mainplot and row spacing as subplot. In the NA x RS experiment, five N levels were

used: 0, 60, 90, 120 and 150 kg ha-1. For NS x RS, a total fixed rate of 100 kg ha-1 was applied,

but under five different timing of application in percent: 0-30-30-30-10 (NS1), 0-20-50-30-0

(NS2), 0-20-30-50-0 (NS3), 23-23-29-25-0 (NS4), and 18-0-29-43-10 (NS5) at the following

schedules: 0 DAE (basal), 10-14 DAE, 30-35 DAE, 45-50 DAE and 60 DAE. Three row spacings

were used in both experiments: 25 cm (RS25), 30 cm (RS30), and 35 cm (RS35). Results showed

that under rainfed condition, about 3.1 to 4.9 t ha-1 of grain yields of aerobic rice can be obtained

with addition of 60-150 kg N ha-1. Grain yields increased with higher application of nitrogen

fertilizers, but beyond N level of 90 kg ha-1, the risk of lodging may pose a problem especially in

the wet season. N-split treatments did not generally affect grain yields, and the common practice

of three to four split applications of N-fertilizer at major crop growth stages of flooded lowland rice

showed is also applicable for aerobic under rainfed condition. Grain yields were not affected by

differences of row spacing. Although the number of panicles per square meter in the narrow row


Page | 132

spacing (25 cm) was significantly higher than in the wider spacing (35 cm), this difference was

compensated for by the significantly higher number of spikelet per panicle in the wider spacing.

The percent filled spikelet was also slightly higher in the wider spacing than in the narrow spacing.

We did find significant effect of row spacings on percent lodging. While the optimum combination

of N level and row spacing on grain yield and bending resistance of aerobic rice was not identified,

results suggested that farmers may grow aerobic rice using row spacing within 25-35 cm as long

as this row spacing may not interfere with other cultural management practices.

Limeng Zhang, Shan Lin, Hongbin Tao, Changying Xue, Xiaoguang Yang, Dule Zhao,

B.A.M. Bouman, Klaus Dittert (in prep 2008), Response of yield determinants and dry

matter translocation of aerobic rice to N application on two soil types

Abstract

At the background of increasing water scarcity, a new type of water-saving rice, aerobic rice has

been developed recently. Little is known about the crop performance and nutrient uptake in

response to nitrogen management and soil types such as lowland and upland soils. Therefore, in

2005-2006, field experiments were conducted with aerobic rice cv. Han Dao 297 (Hereafter,

HD297) under different N fertilizer levels and irrigation regimes on a traditional lowland soil

(Shangzhuang site) and an upland soil (CAU site) near Beijing, North China. The nitrogen rates

were 0 kg N ha-1 (N0), 75 kg N ha-1 (N1) and 150 kg N ha-1 (N2), split applied at a ratio of 3:4:3

given at sowing, tillering and booting stage. The irrigation management was based on the soil

water potential at 15 cm depth maintaining soil moisture at around -20 kPa through the entire

growing season (W1), at -40 kPa from emergence to PI, at -20 kPa from PI to flowering and at -40

kPa after flowering (W2), and at -60 kPa over the entire season (W3). Aerobic rice HD297 yielded

3.1 to 5.3 ton ha-1 as influenced by N fertilizer rates and site conditions. At Shangzhuang site in

2005, HD297 combined a high aboveground biomass (15.6 t dry matter ha-1) with a low harvest

index (averaged 0.30) at the seeding rate 135 kg/ha. Little effect of nitrogen on grain yield was

observed. In 2006, pre-anthesis dry matter production, tiller numbers and LAI clearly increased

with increasing N fertilizer rates at both sites. However, during grain filling, differences in dry

matter were only small, and in some cases, differences in pre-anthesis dry matter were

compensated. Dry matter translocation from vegetative organs into grains increased with N

fertilizer rates, but the mean translocated dry matter was relatively low, with only 1.08 t DM ha-1.

Dry matter translocation efficiency ranged from 3.7-34.9 %. And the contribution of pre-anthesis

assimilates to grain was 10.9-65.7 % differed with N supply. Low dry matter translocations

associated with a sharp decrease in LAI after flowering. And poorer percent filled grains with

higher N supply were exhibited. On average across treatments, 13 % greater grain yield was

produced at Shangzhuang site than at CAU site. Higher dry matter production and greater

spikeltets per m2 at Shangzhuang site contributed to the higher grain yield compared to CAU site.


Page | 133

Lixiao Nie, Shaobing Peng, Bas A. M. Bouman, Jianliang Huang, Kehui Cui, Romeo M.

Visperas, and Hong-Kyu Park, 2007. Solophos fertilizer improved rice plant growth in

aerobic soil. Journal of Integrated Field Science 4: 11-16.

Abstract

Yield decline of continuous monocropping of aerobic rice is the major constraint to the wide

adoption of aerobic rice technology. This study was conducted to determine if solophos fertilizer

could be used to reverse the yield decline of this cropping system using pot and micro-plot

experiments. The soil for the pot experiment was collected from a field where aerobic rice has

been grown continuously for 11 seasons at the IRRI farm. Four rates (4, 6, 8, and 10 g pot-1) of

solophos application were used in the pot experiment. Micro-plots (1 × 1 m) were installed in the

field experiment where the 12th-season aerobic rice was grown. Treatments in the micro-plots

were with and without additional solophos application. Solophos rate was 4,407.5 kg ha-1 which

was equivalent to 10 g solophos pot-1 used in the pot experiment. An improved upland variety,

Apo, was used for both pot and micro-plot experiments. Application of solophos significantly

increased plant height, stem number, leaf area, chlorophyll meter reading, root dry weight, and

total biomass in the pot experiment. The growth enhancement by solophos application was also

observed in the micro-plot experiment under the field conditions. Photosynthetic rate and spikelet

number per m2 were increased by solophos application in the micro-plot experiment. Although the

mechanism of growth promotion by solophos application is not clear, this study suggested that

solophos application can be used as one of crop management options that could minimize the yield

decline of continuous monocropping of aerobic rice.

Lixiao Nie, Shaobing Peng, Bas A.M. Bouman, Jianliang Huang, Kehui Cui, Romeo M.

Visperas, Hong-Kyu Park, 2007. Effect of oven heating of soil on vegetative plant growth

of aerobic rice. Plant and Soil 300: 185-195.

Abstract

“Aerobic rice” system is the cultivation of nutrient-responsive cultivars in nonflooded and

nonsaturated soil under supplemental irrigation. It is intended for lowland areas with water

shortage and for favorable upland areas with access to supplementary irrigation. Yield decline

caused by soil sickness has been reported with continuous monocropping of aerobic rice grown

under nonflooded conditions. The objective of this study was to determine the growth response of

rice plant to oven heating of soil with a monocropping history of aerobic rice. A series of pot

experiments was conducted with soils from fields where rice has been grown continuously under

aerobic or anaerobic (flooded) conditions. Soil was oven heated at different temperatures and for

various durations. Plants of Apo, an upland variety that does relatively well under the aerobic

conditions of lowland, were grown aerobically without fertilizer inputs in all six experiments. Plants

were sampled during vegetative stage to determine stem number, plant height, leaf area, and


Page | 134

total biomass. Heating of soil increased plant growth greatly in soils with an aerobic history but a

relatively small increase was observed in soils with a flooded history as these plants nearly

reached optimum growth. A growth increase with continuous aerobic soil was already observed

with heating at 90 °C for 12 hours and at 120 °C for as short as 3 hours. Maximum plant growth

response was observed with heating at 120 °C for 12 hours. Leaf area was most sensitive to soil

heating, followed by total biomass and stem number. We conclude that soil heating provides a

simple and quick test to determine whether a soil has any sign of sickness that is caused by

continuous cropping of aerobic rice.

Lixiao Nie, Shaobing Peng, Bas A.M. Bouman, Jianliang Huang, Kehui Cui, Romeo M.

Visperas, and Jing Xiang, (in preparation 1), Alleviating soil sickness caused by aerobic

monocropping: Responses of aerobic rice to nutrient supply, to be submitted to Field

Crops Research.

Abstract

Yield decline is a major constraint in the adoption of continuous cropping of aerobic rice. The

causes of the yield decline in the continuous aerobic rice system are still unknown. The objective of

this study was to determine if nutrient application can mitigate the yield decline caused by

continuous cropping of aerobic rice. Micro-plot experiment was conducted in 2005 dry season (DS)

in a field where aerobic rice has been grown continuously for eight seasons from 2001 DS at the

International Rice Research Institute (IRRI) farm. Pot experiments were done with the soil from

the same field where the micro-plot experiment was conducted and aerobic rice has been grown

continuously for 10 seasons. Apo, an upland rice variety, was grown under aerobic conditions with

different nutrient inputs in field and pot experiments. The micro-plot experiment showed that

micronutrients had no effect on plant growth under continuous aerobic rice cultivation but the

combination of N, P, and K mitigated the yield decline of continuous aerobic rice. A series of pot

experiments studying the individual effects of nutrients indicated that N application improved plant

growth under continuous aerobic rice cropping, while P, K, and micronutrients had no effect.

Increasing the rate of N application from 0.23 to 0.90 g per pot in the continuous aerobic rice soil

increased the vegetative growth parameters, chlorophyll meter readings, and aboveground N

uptake consistently. Our results suggested that N deficiency due to poor soil N availability or

reduced plant N uptake might cause the yield decline of continuous cropping of aerobic rice.

Lixiao Nie, Shaobing Peng, Bas A.M. Bouman, Jianliang Huang , Kehui Cui, Romeo M.

Visperas, Jing Xiang, (in preparation 2), Alleviation of Soil Sickness Caused by aerobic

monocropping: Responses of rice plant to various nitrogen sources, to be submitted to

Plant and Soil.

Abstract


Page | 135

Yield decline of continuous cropping of aerobic rice is the major constraint to the wide adoption of

aerobic rice technology. This study was conducted to examine the differences in plant growth

responses to different N sources applied in the continuous aerobic rice soil. Soils for pot

experiments were collected from two adjacent fields at the International Rice Research Institute

(IRRI) farm: an aerobic field where aerobic rice has been grown continuously for 11 seasons from

2001 dry season (DS) and a flooded field where flooded rice has been grown continuously. The

results showed that application of additional N significantly increased the plant growth and grain

yield of aerobic rice in continuous aerobic rice system. Among the nitrogen fertilizers tested,

ammonium sulfate was the most effective in improving plant growth under the continuous aerobic

soil. The plant growth, grain yield, total biomass, N content in grains and aboveground N uptake of

plants fertilized with ammonium sulfate were relatively higher than those fertilized with urea, when

the N rate was above 0.6 g N pot-1. Furthermore, the differences in these parameters between

ammonium sulfate and urea applications increased as the N rate increased. Additional pot

experiment using the continuous aerobic soil was conducted to examine the effect of ammonium

sulfate in Apo, IR80508-B-57-3-B, and IR78877-208-B-1-2. Ammonium sulfate consistently and

significantly improved plant growth in the three genotypes. Our experiments suggested that

ammonium sulfate may be used to mitigate the yield decline caused by continuous cropping of

aerobic rice and that it is possible to reverse the yield decline by using of N efficient genotypes in

combination of improved N management strategies.

Xiao Qin Dai, Hong Yan Zhang, J. H. J. Spiertz, Jun Yu, Guang Hui Xie, B. A. M Bouman,

(in prep 2008), Productivity and Resource Use of Aerobic Rice – Winter Wheat Cropping

Systems in the Huai River Basin.

Abstract

Water shortage is threatening conventional irrigated rice production, prompting the introduction of

water-saving rice production systems in China. Considerably savings on irrigation water are

possible by growing rice in a non-flooded aerobic soil. In aerobic rice systems soil aeration status

and nutrient availability differ from flooded lowland systems. This may affect nutrient availability

within one crop cycle, but also over a cropping sequence. However, the response of aerobic rice to

nutrients and its use efficiency in a cropping sequence has hardly been documented. To study

these responses, a field experiment was conducted with aerobic rice - winter wheat (AR-WW) and

winter wheat - aerobic rice (WW-AR) cropping sequences in the Huai River Basin, China. Fertilizer

treatments comprised of a standard NPK dose, a PK dose (N omission),a NK dose (P omission), a

NP dose (K omission) and a control with no fertilizer input. Omission of N reduced yield by 0.5 ~

0.8 Mg ha-1 for aerobic rice and 2.3 ~ 4.3 Mg ha-1 for winter wheat. The yield loss was less when

only P or K was omitted, indicating that N was the most limiting nutrient for both crops. For the

whole cropping system (aerobic rice + winter wheat), grain yield decreased by 44, 11 and 7% for

N, P and K omission in the AR-WW sequence and by 30, 6 and 0% for N, P and K omission in the

WW-AR sequence, indicating that the cumulative effects of N, P or K omission on yield were

greater in the AR-WW than the WW-AR sequence. Generally, omissions of N, P and K decreased


Page | 136

the nutrient concentrations of various plant parts and the total nutrient uptake of aerobic rice and

winter wheat. Nitrogen, P and K concentration of aerobic rice in WW-AR and of winter wheat in AR-

WW sequence were respectively lower than those in another corresponding sequence, indicating

that nutrient depletion by the preceding crops further decreased the nutrient concentration of the

following crops. Nutrient uptake by aerobic rice and winter wheat was significantly influenced by

the fertilizer treatments and differences in uptake were associated with plant biomass. Aerobic rice

and winter wheat were more sensitive to N than to P or K omission, indicating that N was the most

important limiting nutrient. Furthermore, aerobic rice responded less to nutrient omissions than

winter wheat, because the latter had a higher nutrient demand. The highest nutrient use

efficiencies were associated with a low nutrient availability and low yields. The challenge should be

to improve crop productivity and resource use efficiency simultaneously not only for one individual

crop, but for the whole cropping system. To combine high nutrient use efficiencies and

productivity, an appropriate well-balanced fertilizer management is required.

Xue Changying, Yang Xiaoguang, Bouman B.A.M., Deng Wei, Zhang Qiuping, Yan

Weixiong, Zhang Tianyi, Rouzi Aji, Wang Huaqi,Wang Pu, (submitted), Effects of

irrigation and nitrogen on the performance of aerobic rice in northern China. Journal of

Integrative Plant Biology.

Abstract

Aerobic rice is a new production system in which specially-developed varieties are grown under

nonflooded, nonpuddled, and nonsaturated soil conditions. Insight is needed into water and

fertilizer N response and water by N interaction to develop appropriate management

recommendations. In 2003-2004, irrigation x N experiments were done near Beijing using variety

HD297. Water treatments included four irrigation levels, and N treatments included different

fertilizer N application rates and different number of N splits. The highest yields were 4.5 t ha-1

with 688 mm of total (rain plus irrigation) water input in 2003 and 6.0 t ha-1 with 705 mm of water

input in 2004. Because of quite even distribution of rainfall in both years, the four irrigation

treatments did not result in large differences of soil water conditions. There were few significant

effects of irrigation on biomass accumulation, but yield increased with total amount of water

applied. High yields coincided with high harvest index and high percentage of grain filling. The

application of fertilizer N either reduced biomass and yield or kept it at the same level as 0 N and

consistently reduced percentage grain filling and 1000-grain weight. There were no or inconsistent

interactions between water and N. With the highest water application, five splits of N gave higher

yield than three splits, whereas three splits gave higher yield than five splits with lower water

applications. High yields with 0 N were probably caused by frequent overfertilization in the past,

leading to high levels of indigenous soil N supply. A longer-term (over three years) experiment

may be needed to quantify the N response of aerobic rice and how much N fertilizer can be saved

in N-saturated soils.


Page | 137

Xue Changying, Yang Xiaoguang, B.A.M. Bouman, Deng Wei, Zhang Qiuping, Yan

Weixiong, Zhang Tianyi, Rouzi Aji, Wang Huaqi, (submitted), Optimizing yield, water

requirements, and water productivity of aerobic rice for the North China Plain, Irrigation

Science.

Abstract

Water resources for agricultural are rapidly declining in the North China Plain because of increasing

industrial and domestic use and because of decreasing rainfall resulting from climate change.

Water-efficient agricultural technologies need to be developed. Aerobic rice is a new crop

production system in which rice is grown in nonflooded and nonsaturated aerobic soil, just like

wheat and maize. Although an estimated 80,000 ha are cultivated to aerobic rice in the plain,

there is little knowledge on obtainable yields and water requirements to assist farmers in

improving their management. We present results from field experiments with aerobic rice variety

HD297 near Beijing, 2002-2004. The crop growth simulation model ORYZA2000 was used to

extrapolate the experimental results to different weather conditions, irrigation management, and

soil types. We quantified yields, water inputs, water use, and water productivities. On typical

freely-draining soils of the North China Plain, aerobic rice yields can reach 6-6.8 t ha-1, with, on

average some 477 mm rainfall and 112-320 mm of irrigation water. The irrigation water can be

supplied in 2-4 applications and should aim at keeping the soil water tension in the root zone

below 100-200 kPa. Under those conditions, the amount of water use by evapotranspiration is

458-483 mm. The water productivity with respect to total water input (irrigation plus rainfall) is

0.89-1.05 g grain kg-1 water, and with respect to evapotranspiration, 1.28-1.42 g grain kg-1 water.

Drought around flowering should be avoided to minimize the risk of spikelet sterility and low grain

yields. The simulations suggest that, theoretically, yields can go up to 7.5 t ha-1 and beyond.

Further research is needed to reveal whether the panicle (sink) size is large enough to support

such yields and/or improved management is needed.


Page | 138

Appendix B: Ten Frequently Asked Questions and answers

What is aerobic rice?

Aerobic rice is a production system in which specially developed, input-response rice varieties with

“aerobic adaptation” are grown in well-drained, nonpuddled, and nonsaturated soils without

ponded water, with a management system aiming at yield levels of 4-6 t ha-1 (and possibly

beyond). A nonsaturated soil is also called an “aerobic soil”.

What are aerobic rice varieties?

Varieties adapted to aerobic management systems require the ability to maintain rapid growth in

soils with moisture content at or below field capacity. They share this ability with traditional upland

rice varieties, which usually have deep root systems and tolerate water stress at both the

vegetative and reproductive stages. However, varieties for aerobic production systems also need

to be able to produce yields of 4-6 t ha-1 under favorable conditions. Traditional upland varieties,

which are usually low-tillering, tall, and have a low harvest index, rarely achieve yields higher than

3 t ha-1 even under the most favorable conditions. Achieving high yields under aerobic soil

conditions requires new varieties of “aerobic rice” that combine the drought-resistant

characteristics of upland varieties with the high-yielding characteristics of lowland varieties.

Aerobic rice varieties combining high yield potential with tolerance to aerobic soil

conditions have usually been derived from breeding programs in which varieties are developed and

evaluated under aerobic soil conditions and with fertilizer applications sufficient for a 4-6 t ha-1

yield target. The first generation of varieties that performed well in a wide range of aerobic rice

environments (e.g. IR55423-01 (“Apo”) and UPLRI-5 from the Philippines, B6144-MR-6-0-0 from

Indonesia, and CT6510-24-1-2 from Colombia) were developed in upland rice breeding programs.

They were often derived from crosses between indica and tropical japonica parents, whereas

traditional upland varieties are usually derived from the aus or tropical japonica germplasm

groups. Some aerobic rice breeding programs, (notably that of the China Agricultural University in

Beijing), also have developed successful varieties by crossing high-yielding lowland rice varieties

with traditional upland types. In northern China, new elite aerobic varieties were released in the

late 1990s such as Han Dao 277, Han Dao 297 and Han Dao 502, with yield potentials of up to 6.5

t ha-1.

What is the difference between aerobic rice and upland rice?

Upland rice is grown in rainfed, naturally well-drained soils with bunded or unbunded fields without

surface water accumulation. The general perception about the upland environment is that it is

drought-prone, usually sloping land with erosion problems, and has soils with both poor physical

and chemical properties. Farmers in these environments are among the poorest and usually can

not afford to apply (many) external inputs such as fertilizers. Upland rice varieties are mostly

grown as a low-yielding subsistence crop to give stable yields under the adverse environmental

conditions of the uplands. Upland rice varieties are drought tolerant, but have a low yield potential


Page | 139

and tend to lodge under high levels of external inputs such as fertilizer and supplemental

irrigation.

The aerobic rice system is targeted at more favorable environments (see below) where

farmers can afford to buy external inputs such as fertilizers and have access to supplementary

irrigation if rainfall is not sufficient. Achieving high yields under relatively favorable aerobic soil

conditions requires new varieties of “aerobic rice” that combine the drought-resistant

characteristics of upland varieties with the high-yielding characteristics of lowland varieties. In

essence, aerobic rice can be seen as “favorable” or “high yielding” upland rice. The reason for the

introduction of a new term was the need to dissociate the envisioned relatively high-yielding

production system from the general perception of extremely harsh and unfavorable conditions of

“the uplands”.

Why aerobic rice?

There are three main driving forces for aerobic rice:

1. The increasing realization that not all “uplands” are “unfavorable”, in the sense that certain

uplands may possess soils with good water-holding capacity and high fertility, that they

are not always sloping land, that rainfall may be sufficient for a “decent’ crop growth, and

that sometimes investments can be made to improve the quality of the uplands. An

example of the latter is the terracing of slopes in the hilly and mountainous regions in

Yunnan, China, and in North Vietnam. Aerobic rice is seen as a relatively high-yielding

production system that optimally exploits the resources available. Other examples of

“favorable” uplands are the flat Cerrado region in Brazil, and the North China Plain where

aerobic rice is introduced in typical high-yielding upland cropping systems (such as maize,

cotton, soybean).

2. The increasing awareness that many areas in the so-called “rainfed lowlands” don’t receive

enough water to keep the rice fields predominantly flooded. Rainfed lowlands are often

characterized by slightly undulating topography with differences in elevation of a few

meters across a toposequence of a few hundred meters only. Because of this topography,

however, fields at the top of a toposequence often have deep groundwater tables, more

coarse-textured soils, and more runoff and seepage losses. The soils in these fields are

often dominantly aerobic and hence an ideal target domain for the system of aerobic rice.

3. The increasing water scarcity in irrigated lowlands. The causes for water scarcity are

diverse and location-specific, but include decreasing resources (e.g., falling groundwater

tables, silting of reservoirs), decreasing quality (e.g., chemical pollution, salinization),

malfunctioning of irrigation systems, and increased competition from other sectors such as

urban and industrial users. In extreme cases, water scarcity can be so severe that farmers

can not maintain flooded conditions in their fields for even a small part of the growing

season, and rice fields are not ponded and saturated with water anymore. However,

irrigation water availability is still sufficient for supplementary irrigation to keep the soil

water content around field capacity. Under such conditions, lowland rice can not be grown


Page | 140

anymore, and aerobic rice becomes a suitable alternative along with upland crops

(diversification).

Where aerobic rice?

Aerobic rice can be found, or can be a suitable technology, in the following major rice-growing

environments:

1. So-called “favorable uplands” (see FAQ 4: Why aerobic rice?): areas where the land is flat

(or terraced), where rainfall with or without supplemental irrigation is sufficient to

frequently bring the soil water content close to field capacity, where no serious soil-

chemical limitations such as aluminium toxicity or salinity occur, and where farmers have

access to external inputs such as fertilizers. A typical example is in the Cerrado region of

Brazil, where farmers grow aerobic rice in rotation with crops such as soybean and fodder

on large commercial farms with supplemental sprinkler irrigation on an estimated 250,000

ha of flat lands, realizing yields of 3-4 t ha-1. Another example is rainfed aerobic rice grown

in newly-formed terraces in the hills of Yunnan, China, where yields are also typically 3-4 t

ha-1.

2. Fields on upper toposequence locations in undulating so-called “rainfed lowlands”. Quite

often, the soils of such upper fields or terraces are relatively coarse-textured and well-

drained, so that ponding of water only occurs for a limited (or no) part of the growing

season. No widespread examples of aerobic rice in rainfed lowlands are known, but these

upper fields have been proposed as target domain for aerobic rice.

3. Water-short irrigated lowlands (see FAQ 4: Why aerobic rice?): areas where farmers do

not have access to water to keep rice fields flooded for a substantial period of time

anymore. Water shortage can be encountered in tail-end parts of large-scale surface

irrigation systems, in areas where the groundwater has been drawn down so that pumping

costs have become very high, in irrigation systems that receive less and less water

because of redirected use (cities, industry) or because of reduced stream flow in rivers. A

good example is the North China Plain where aerobic rice is grown on about 80,000 ha

with supplemental irrigation.

Beside these typical rice-growing environments, aerobic rice can also be found in traditionally non-

rice growing areas. Again in the North China Plain, farmers are experimenting with aerobic rice as

a means of crop diversification in areas where traditionally maize is the dominant crop.

Aerobic rice can be found in tropical and in temperate climates. Most advances in

developing aerobic rice systems, and in adoption by farmers, have been made so far in China and

Brazil.

How to manage aerobic rice?

Aerobic rice is basically managed like a wheat or a maize crop. The usual establishment method is

dry direct seeding. Before sowing, the land should be dry prepared by ploughing and harrowing to

obtain a smooth seed bed. Seeds should be dry seeded at 1-2 cm depth in heavy (clayey) soils

and 2-3 cm depth in light-textured (loamy) soils. Optimum seeding rates still need to be


Page | 141

established but are probably in the 70-90 kg ha-1 range. In experiments so far, row spacings

between 25 and 35 cm gave similar yields. The sowing of the seeds can be done manually (eg

dibbling the seeds in slits opened by a stick or a tooth harrow) or using direct seeding machinery.

An alternative establishment method is transplanting, where seedlings are transplanted into wet

soil that is kept around saturation for a few days to ease transplanting shock. Subsequently the

fields dry out to field capacity and beyond. This method of crop establishment can only be done in

clay soils with good water-holding capacity.

An aerobic rice crop attaining 4-6 t ha-1 yields obtains many of its required nutrients from

the soil. But this “indigenous supply” of nutrients is typically not sufficient to meet all the nutrient

needs, and fertilizers will need to be applied. Site-specific knowledge about the indigenous

nutrients supply is the starting point for formulating fertilizer recommendations. The site-specific

nutrient management (SSNM) approach (www.irri.org/irric/ssnm) can be used to determine the

need for supplemental nutrients in the form of fertilizers and the optimal management of

fertilizers. A useful tool to assist in the application of nitrogen (N) fertilizer is the Leaf Color Chart

(LCC; part of the SSNM approach). In the absence of trained extension personnel in SSNM and

LCC, an amount of 70-90 kg N ha-1 could be a useful starting point (to be subsequently optimized).

Instead of basal application of the first N split, the first application can best be applied 10-12 days

after emergence to minimize N losses by leaching (the emerging seedling can’t take up N so fast,

so it will easily leach out). Moreover, basal application of N also promotes early weed growth.

Second and third split applications of N may be given around active tillering and panicle initiation,

respectively. Dry, aerobic, soil can reduce the indigenous supply of phosphorus (P), hence the

application of fertilizer P can be more critical for aerobic rice than for conventional flooded lowland

rice. On acid soils, aerobic rice will likely be less prone to zinc deficiency than flooded lowland rice;

but on high pH soils with calcium carbonate, the reverse may be true.

If the crop is grown in a dry season, a light irrigation application (say 30 mm) should be

given after sowing to promote emergence. Subsequent irrigation applications depend on the

rainfall pattern, the depth of groundwater, and on the availability and/or cost of irrigation water.

Irrigation can be applied by any means as used for upland crops: flash flood, furrow, or sprinkler.

Rice that is not permanently flooded tends to have more weed growth and a broader weed

spectrum than rice that is permanently flooded. To control weeds, the use of pre- or post-

emergence herbicides is recommended when the weed pressure is high, plus additional manual or

mechanical (inter-row cultivation) weeding in the early phases of crop growth.

Is aerobic rice rainfed or irrigated?

Like wheat or maize, aerobic rice can be rainfed, supplementary irrigated, or fully irrigated. With

groundwater tables below the root zone, a suitable total amount of water supply by rainfall and/or

irrigation is probably 600-800 mm over the growing season. With deep groundwater tables and

less than 400 mm, the typical traditional upland rice system with sturdy and drought-resistant

upland varieties is more suitable. When subsurface hydrology, soil type, and rainfall (and/or

irrigation) combine to create predominantly flooded or saturated soil conditions throughout the

growing season, the typical lowland rice production system is more suitable. With clayey soils and

groundwater tables below the root zone, one would need typically 1000 mm or more to maintain


Page | 142

predominantly saturated soil conditions. However, with groundwater tables within the 20-cm root

zone (as occurs in many typical irrigated lowland rice environments), as little as 400 mm can

already maintain predominantly saturated soil conditions or flooding. The exact “transition zones”

between upland rice and aerobic rice, and between aerobic rice and lowland rice production

systems is therefore quite site-specific.

The optimum soil water condition for aerobic rice is around field capacity. If rainfall is

insufficient to frequently restore water contents in the soil to field capacity, irrigation can be

applied if water resources are available. Irrigation can be applied through flash-flooding, furrow

irrigation (or raised beds), or sprinklers. Unlike flooded rice (lowland rice), irrigation - when

applied – is not used to flood the soil but to just bring the soil water content in the root zone up to

field capacity. The amount of irrigation water should match evaporation from the soil and

transpiration by the crop (plus any application inefficiency losses). In lowland rice, the amount of

irrigation water should match the same water flows, plus the losses by seepage and percolation.

Is aerobic rice a “mature” technology?

Aerobic rice can be considered quite a mature technology in temperate and subtropical

environments such as northern China and Brazil, where the areas of aerobic rice are estimated at

80,000 ha and 250,000 ha, respectively. In both countries, breeding programs since the 1980s

have resulted in the release of several high-yielding “aerobic rice” varieties. On-farm yield levels

seem to lie around 3-4 t ha-1, but yields of up to 6 t ha-1 have been recorded as well. Current

research focuses on the development of improved management systems and on breeding further

improved varieties.

Tropical aerobic rice systems are still very much in the research and development phase.

More research is especially needed to breed high-yielding aerobic rice varieties with sufficient

aerobic adaptation and to develop sustainable management systems. Without ponded water, rice

production is less sustainable than under flooded (lowland) conditions, and typical problems come

up that occur in upland crops (see FAQ 10: How sustainable is aerobic rice?). In general,

sustainability seems to be more of a problem in tropical areas than in temperate areas such as

northern China. Aerobic rice should not be grown consecutively on the same piece of land, and –

depending on the cropping history and soil type – low yields can even occur on fields cropped to

aerobic rice the very first time.

How about aerobic rice and conservation agriculture?

With aerobic rice, practices of conservation agriculture, such as mulching and minimum tillage as

practiced in upland crops, become available to rice farmers as well. In the Indo-Gangetic Plain,

farmers are experimenting with minimum tillage practices and permanent raised beds in the rice-

wheat system. Pioneering research and development work is being done by the Rice Wheat

Consortium (http://www.rwc.cgiar.org/index.asp).

Various methods of mulching (e.g., using dry soil, straw, and plastic sheets) are being

experimented with in aerobic rice systems in China. In hilly areas in Shiyan, Hubei Province in

China, farmers on an estimated 6000 ha are adopting the use of plastic sheets to cover rice fields

in which the soil is kept just below saturation. The proclaimed advantages are: earlier crop


Page | 143

establishment (rice is established in early spring when temperatures are still low, and the plastic

sheet increases the soil temperature), higher yields, less weed growth, and less water use

(important during dry spells). The left-over plastic after harvest may cause environmental

degradation if not properly taken care of.

How sustainable is aerobic rice?

Given assured water supply, lowland rice fields are extremely sustainable and able to produce

continuously high yields, even under continuous double or triple-cropping a year. Flooding of rice

fields has beneficial effects on soil acidity (pH), soil organic matter buildup, phosphorus, iron, and

zinc availability, and biological N fixation that supplies the crop with additional N. When fields are

not continuously flooded, such as in aerobic rice, these beneficial effects gradually disappear. A

change from flooded to aerobic soil conditions may decrease the soil organic matter content,

decrease the soil pH, and decrease the availability of phosphorus, iron, and – on calcareous soils -

zinc. Also, problems with micro-nutrient deficiencies have been reported. If field were cropped to

rainfed rice with alternate periods of flooding and dry soil, or if fields were previously cropped to

upland crops, then the introduction of aerobic will have fewer consequences for these sustainability

parameters.

There are indications that soil-borne pests and diseases such as nematodes, root aphids,

and fungi occur more in aerobic rice than in flooded rice, especially in the tropics. The current

experience is that aerobic rice should not be grown continuously on the same piece of land each

year (as can be successfully done with flooded rice) without yield decline. Suitable crop rotations

need to be identified, but will be site-specific and responsive to markets.

Current research focuses on determining the causes of yield decline under continuous

cropping (biotic, abiotic), on developing “resistant” varieties, on developing suitable management

options such as crop rotation, and on developing integrated weed management practices.


Page | 144

Appendix C: Detailed recommendations for further research

Liz Humphreys, CPWF Theme Leader, (November 2007)

The development of systems of aerobic rice for tropical and temperate environments (STAR) is

very important, with huge potential benefits to hundreds of millions of people in terms of food

security, livelihoods and the environment.

There has been a lot of progress in the development of STAR – from understanding the genetics

and mechanisms of drought tolerance, to improved varieties and how to manage them for optimal

land and water productivity.

However there is still much to do – the work has hardly begun apart from in China and to a lesser

degree in Philippines. Aerobic rice may be beneficial in many other regions, especially in South

Asia, in both water scarce irrigated areas and favourable rainfed uplands.

The challenge will be to identify the most important research and development priorities from the

plethora of possible activities. Some of the areas which I think need further emphasis are:

1. Impact of widespread adoption on regional hydrology and ecology

There is a lot of publicity about and emphasis on irrigation water savings with aerobic rice, but

almost no consideration of total water savings, which are likely to be much less than irrigation

water savings. Irrigation water savings are important, meaning increased efficiency of resources

such as energy for pumping groundwater. However some or much of the savings may be due to

reduced deep drainage into the groundwater, or reduced runoff. Where groundwater and runoff

can be used elsewhere in the system (recycling), this is not a water saving.

There need to be:

- further studies quantifying components of the water balance for aerobic rice for a few case

study situations – field experimentation and application of crop models

- education of researchers and extensionists on irrigation water savings versus system water

savings (reduced depletion to sinks), so that water resource managers and policy makers

are aware of the absolute water requirements of aerobic rice systems

- regional scale studies of the impacts of widespread adoption of aerobic systems on the

regional hydrology and ecology

2. Start developing aerobic systems incorporating “conservation agriculture”

approaches

Temperate and sub-tropical drill seeded aerobic rice systems are likely to be well-suited to take

advantage of the benefits of conservation agriculture (reduced tillage, mulching, crop rotation,

brown manuring). Such an approach would help conserve soil moisture, increase soil fertility and


Page | 145

control diseases. It may be beneficial to develop links with individuals/groups with expertise in

conservation agriculture, including those from PN12 in China and Mexico (CIMMYT) and several

groups with experience in rice-wheat systems in South Asia.

Such work needs to be long term, and could be incorporated in the long term trials suggested

below.

3. Long term trials with multi-disciplinary teams

There are issues about the sustainability of aerobic rice systems. In some situations, yields decline

after a couple of crops. Long term trials are needed across a range of situations (soil, climate)

There are also many potential causes of aerobic rice not reaching potential yields. A

multidisciplinary approach involving teams of scientists is needed to explain crop and system

performance. Such teams would include breeders, agronomists, crop physiologists, soil chemists,

soil biologists, water scientists, economists etc.

AK Singh’s group at IARI, India have a couple of very good examples of long term trials involving

multidisciplinary teams of researchers, and perhaps some things to learn about how to do this

successfully (i.e. everyone actually contributes).

4. Approaches to achieve rapid dissemination and widescale adoption

Need to identify and develop approaches to achieve rapid dissemination and widescale adoption –

opportunity to build on approaches introduced by CPWF to help achieve this?

5. Systematic site characterization and monitoring

The amount of characterization and monitoring will vary depending on the purpose, but protocols

tailored to some key purposes need to be developed and disseminated to explain results, synthesis

of findings across locations and develop generic understanding, and extrapolate findings across

locations.

In addition to determination of baseline properties and in-season monitoring, a system of archiving

soil samples may be useful for some situations to enable post experiment determination of soil

properties as new knowledge and ideas arise.

e.g. protocols for breeders undertaking variety evaluation should include monitoring of watertable

depth (simple), weather (rain, evaporation), soil type (texture of soil layers to ~0.5 m)

e.g. protocols for agronomists/nutritionists undertaking N management trials should include

selection of sites that will be responsive to N (based on site history, soil test)

e.g. protocols/proforma for rigorous financial analysis

Grant Singleton, IRRC coordinator (November 2007)

From an ecologist’s perspective, who is a newcomer to aerobic rice, I provide the following

comments:


Page | 146

1. I strongly agree with Liz’s comment that we need a landscape level study of the benefits of

aerobic rice with respect to water savings In particular; we need to take into account the water

that percolates back into the irrigation system or into streams and rivers (and therefore

contributing directly to their health). I see this as a challenging but exciting research question.

2. As highlighted by Achim, we need a better understanding of factors that influence farmers in

adopting/adapting an aerobic rice cropping system. Socio-economic studies have been conducted

in the Philippines, and, to a lesser extent, in China. Indeed, I look forward to seeing the report

from Dr. Shijun Ding from Zhongnan Univ on the impact assessment of aerobic rice in North China.

Once this report is available then it would be timely to review the socio-economic lessons learned

from the Philippines and China.

3. We need to be able to anticipate and then develop management practices for the impact of

pests, particularly weeds, but also insects and rodents at a scale larger than a single field. I

discuss nematodes in point 4.

4. The research on nematodes is progressing nicely and is impressive given the person power

involved, but we need to understand further the mechanism of the host-parasite interaction. There

is much that can be learned from the literature on host-parasite interactions in vertebrates,

particularly the need to understand the degree of aggregation of parasites in the landscape/hosts

(typically a negative geometric distribution in mammals) and the importance to quantify the

intensity of infection rather than just the prevalence of infection. For example, is there a threshold

level of parasite load (=intensity) before the plant defenses are unable to cope or is it a linear

response?

5. To develop a better understanding of crop rotations in managing possible yield declines

(including the impact of nematodes and weeds) of aerobic rice.

Achim Doberman, IRRI program Leader (November 2007)

1. Key drivers for aerobic rice include the need to locally produce rice under a lack of water,

rising costs of labor and other inputs, and risk of waterlogging when upland crops are grown as

alternative crops on flat land with poor drainage. The principal target areas for aerobic rice

seem reasonably well understood (e.g., water-deficit lowlands with 500-900 mm rain during

the growing season, lowlands and uplands with flooding risk, and sloping uplands with

potential for intensification on terraced land). A first regional analysis suggests relatively large

areas with potential for aerobic rice in Asia, but also Africa and South America. In Asia, the

farm surveys done so far mostly show average aerobic rice yields in the 3.5-4.5 t/ha range

and net returns that are mostly on par or somewhat below those of irrigated rice. But there is

large variability in its performance and there are also failures. Overall, I think that much

greater economic potential exists if we can further raise the genetic yield potential under

aerobic cultivation conditions and fine-tune soil and crop management practices. Wide-scale

adoption will probably require stable yields in the 5 t/ha range. I also see many parallels to the

issues we discuss and start to address in our emerging research on diversification of irrigated

rice systems towards rice-maize, for example.


Page | 147

2. Some confusion about terminology and cultural practices persists, particularly in China.

Aerobic rice can mean quite different things in terms of cultivars used, hydrological conditions

(groundwater table), number and timing of irrigations, and other crop management practices.

Hence a systematic collection of information on existing practices may form the basis for a

better description and classification (typology) of key aerobic rice systems (combinations of

soil hydrology, crop rotation, cultivar types and cropping practices). This could also contribute

to a more general decision tree for choosing the best cultivar and management options for

maximizing productivity and water productivity in different environments, along the whole

gradient of options for reducing water use in rice. Knowing soils and groundwater tables seems

to be of particular importance for making the right choice. A better typology would also

provide better guidance for breeding targets, target domain characterization and deployment

of germplasm to different areas.

3. I suggest to conduct a more detailed biophysical and socioeconomic analysis of selected larger

target domains, including a strategic assessment of what other alternatives farmers could

pursue. For example, if the alternative is diversification to other crops, what potential (not the

current average production) would those have compared to aerobic rice if both are managed

right?

a. The current household level assessments (e.g., in China) compare aerobic rice with

crops such as maize or soybean, but at yield levels that do not represent achievable

levels for aerobic rice and those other crops. Maize improvement is progressing fast in

China and there are also management practices that could be implemented to reduce

the risk of yield losses of maize or other upland crops due to waterlogging.

b. Some of the household level studies seemed to contain somewhat confusing numbers.

Since the gross margin analysis on one side depends very heavily on yield, measuring

that rather than relying on what farmers report might be a major improvement by

itself.

c. The decision to grow aerobic rice or other crops is one that needs to be analyzed more

systematically, also experimentally. Such an analysis could have strong linkages with

IPSA activities on assessing the potential for rice-maize systems, including jointly

conducted experiments that represent current and (future) optimized management

practices. For example, an irrigated rice farmer in the Philippines can have a number of

options for both dry and west season crops. Should he/she go for aerobic rice in the

DS or WS? Where would maize fit in best? What is less risky and more sustainable? We

need to develop a good framework for answering these questions, including a more

standardized approach for the socioeconomic and impact assessment.

4. With the exception of cooler, long growth duration environments, current yield ceilings of

aerobic rice grown in areas with low water table appear to be around 6 t/ha. Insufficient sink

size and harvest indices of about 0.4 indicate potential for further breeding efforts. To make

progress, breeding may have to become more focused on specific traits related to sink size,

sink-source relations, and root systems.


Page | 148

a. I suggest that we review the management practices in the AYTs of IRRI’s aerobic rice

breeding program to verify that we are not selecting under unrepresentative stress

conditions. Yields presented for AYTs at IRRI seemed quite low, particularly for the first

generation of materials and trials conducted before 2007, when yields were mostly in

the 3-4 t/ha range under so-called “non-stressed aerobic rice” (irrigated when soil

moisture in 15 cm depth is less than 20 kPa). To me, such yields indicate presence of

stresses other than water and we should assure ourselves that those are not related to

location-specific management or other factors. Breeders need to get support from the

soil scientists and agronomists for this. I think our target should be to consistently

achieve DS yields in the 5.5 to 6 t/ha range for good entries in the AYTs.

b. Can simple diagnostic methods for root traits be designed for variety screening?

Although one could argue that there is a relationship between (selecting for)

aboveground growth and root growth, that relationships will also be location-specific

and would not allow us to specifically screen for root health related traits or tolerance

to nematodes. Hence, there could be value in designing doable screening methods for

root traits and those could be included in the final stages of the breeding cycle, to at

least screen the (few) most promising entries in more detail. The same could be done

for physiological selection criteria related to yield potential, but the breeders would

need more support for this.

c. I remain uncertain about the potential of breeding for adaptation to the “soil sickness

syndrom” because for that to be successful we need to gain more insights on what that

soil sickness actually is. To do this right would also require a much better soil and plant

characterization at plot level (because of the patchy nature of “soil sickness”),

including assessment of nematodes (roots)..

5. Agronomists have difficulties to achieve the apparent yield potential of aerobic rice cultivars

and many questions remain unanswered about the causes of yield failures or yield declines. A

major problem in trying to unravel causes of yield failure or yield decline is that such causes

tend to be quite location-specific by nature. Unfortunately, current research on yield decline is

restricted to the IRRI site because no other long-term studies exist outside IRRI.

a. How much research should be invested in understanding the yield decline in aerobic

rice? Arguments in favor are that a scientific understanding is necessary to provide

accurate information on target domains and management practices for aerobic rice.

Arguments against it include (i) the likelihood of site-specific causes of yield decline(s)

and potential lack of a generic, intrinsic factor causing it, and (ii) indications that yield

declines can be avoided by crop rotation or possibly also improved management of

nitrogen (and other nutrients). I suggest addressing this issue as part of a broader

effort that includes a set of carefully designed and managed medium- to long-term

experiments in which integrated management options for aerobic rice are evaluated in

comparison with other principle crop intensification or diversification options (see 3.).

Now is the time to establish those, particularly in key target areas that are or may

soon be undergoing a transition to aerobic rice or other crops.


Page | 149

b. Understanding causes of sudden yield failures (e.g., those observed in the Philippines)

may be the more important challenge. Why, for example, this does not seem to occur

in China is an important question to study. Besides events of extreme nematode

infestations, nutritional problems also seem to sit at the heart of these failures, but

including interactions with biotic stresses that may make aerobic rice plants more

susceptible to such stresses. In addition to the good diagnostic work that has been

done so far, better monitoring of yields, roots and soils and other key characteristics in

major aerobic rice areas may be required, as part of a more strategic assessment (see

3.).

c. It appears that there is much scope for fine-tuning the management of aerobic rice to

get closer to the yield potential claimed by breeders. There may be four areas of

particular importance for closing yield gaps: (i) sustainable cropping systems

(rotations, management and germplasm that avoid biotic stresses and undesirable

changes in soil nutrient supply, including soil pH increases), (ii) fine-tuning of irrigation

timing to avoid water stress in the most critical periods (flowering), (iii) fine-tuning of

N management (timing, N sources, and possibly even N placement), (iv) improved

micronutrient management guidelines (and diagnostics) for aerobic rice. Much of what

we know for flooded rice is probably not directly applicable to aerobic rice. It may be

time to explore SSNM concepts for aerobic rice, including a stronger emphasis on

micronutrients.

6. Continue doing more below-ground work. Major differences in physico-chemical and biological

soil processes exist between aerobic and flooded rice systems that affect the dynamics of

many nutrients, but not many comprehensive studies have been conducted. Gradual changes

in soil chemical properties and behavior may have contributed much to the yield decline under

aerobic rice at IRRI, which appears to be somewhat reversible by applying more N, particularly

acidifying N sources.

a. For some nutrients such as phosphorus, the mechanisms affecting changes in

availability under flooded vs. aerobic conditions are well understood and I see less

need to work on that. Nitrogen and micronutrients are probably the major issues to

focus on, particularly in relation to pH profiles across the root surface to bulk soil

profile, and how all that is affected by different management practices and also cultivar

differences.

b. I think it is of particular importance to understand the major differences in rhizosphere

chemistry and microbiology between aerobic and flooded rice and also genotypic

variation in this. Germplasm adapted to aerobic soil may have greater preference for

mixed NO3/NH4 nutrition and may also possess enhanced nutrient solubilization

mechanisms related to root traits, but this needs to be studied in greater detail. Such

studies must be conducted in real soil, not in solution culture. The sandwich techniques

employed by Guy Kirk in the 1990s can be very valuable for that, but such work

requires dedicated staff/graduate students. We should probably source this out to


Page | 150

partners such as the plant nutrition group at CAU or ARIs, or try to find a Chinese

student/PostDoc to apply for a CSC scholarship. This would have to be pursued now.

c. An interesting question is that of root turnover during the growing season. I’m not

aware of any studies on that in aerobic rice, but it is likely that root turnover differs

from that in a flooded soil, which would have significant impact on microbial processes

near roots. From our work in maize we know that fine root turnover and root exudation

account for the bulk of soil respiration during the growing season and there must be

large differences between a flooded and an aerobic rice crop.

7. There appears to be need for a broader umbrella through which activities can be coordinated

better across countries, between institutions within large countries such as China, and also

between breeders and soil/water/crop scientists. In China, many different groups have started

to work on aerobic rice, including many different breeding programs (and approaches), but

they seem to communicate little with each other and there also seems to be little cooperation

between breeders and soil and crop scientists. We see that often nowadays. Many have access

to their own funding sources, but they and we could probably benefit from working more

closely together, including sharing germplasm and information on cultivation practices and

conducting more integrated breeding-agronomy research programs. Facilitating this is a role

IRRI can play and one that could become a major, broader focus for the next phase of the

STAR project.

cpwf project report · 2016. 8. 2. · taking any action based on the information in this...

Documents